Abstract Book - Neuromodulation Symposium

Transcription

Abstract Book - Neuromodulation Symposium
3rd ANNUAL
MINNESOTA
NEUROMODULATION
SYMPOSIUM
ABSTRACT BOOK
APRIL 17, 2015
THE COMMONS HOTEL
MINNEAPOLIS, MN
neuromodulation.umn.edu
Poster Number: NM-101
Poster Name: Deep Brain Stimulation of the Subthalamic Nucleus Creates an Informational Lesion in
the 6-Hydroxydopamine Rat Model of Parkinsonism
Authorship List: Collin Anderson, Alan "Chuck” Dorval
Institution: University of Utah
Abstract: The death of dopaminergic neurons in the substantia nigra pars compacta leads to motor
symptoms in Parkinson's disease (PD) through changes in electrophysiological activity in neurons of the
basal ganglia. High-­frequency deep brain stimulation (DBS) partially treats these symptoms, but the
mechanisms are unclear. We hypothesized that motor symptoms of PD are associated with increases
in information transmission between the substantia nigra reticulata (SNr) in the ventral basal ganglia
and the ventral anterior thalamus (VA), and that functional DBS treats these symptoms by eliminating
extraneous information transmission between these regions. We tested these hypotheses in a 6­hydroxydopamine-­lesioned rodent model of hemiparkinsonism. In the hemiparkinsonian condition,
information transfer was significantly increased in both the orthodromic and antidromic directions, and
these changes were reversed by functional DBS. There are not known to be significant projections
from VA to SNr, so we interpreted the increase in information in both directions in the
hemiparkinsonian condition as signifying a strengthening of inputs common to both SNr and VA. This
would lead to less independence among informational channels in the basal ganglia, creating an
inability to properly filter extraneous information that creates symptoms. We interpreted that
additional noisy information leads to an inability of the thalamus to properly relay information
regarding healthy motor activity back to cortex, thereby diminishing healthy motor activity, resulting in
some of the hallmark symptoms of PD. In this model, the role of DBS would lie in its regularization of
firing patterns, thereby removing pathological in formation and effectively treating motor symptoms.
Poster Number: NM-102
Poster Name: Mindlock: A Novel Behavioral Syndrome induced by Deep Brain Stimulation in the
Intact and Awake Non-human Primate
Authorship List: Jonathan L. Baker, Xuefeng F. Wei, Jae-Wook Ryou, Christopher R. Butson, Nicholas
D. Schiff, Keith P. Purpura
Institution: Weill Cornell Medical College - Brain and Mind Research Institute
Abstract: Here we report a novel behavioral syndrome, ‘mindlock’, induced by deep brain stimulation
(DBS) of the fasciculus retroflexus (FR) in an intact and behaving non-human primate. The FR is a fiber
bundle composed of medial (MHb) and lateral (LHb) habenula efferents that synapse within midbrain
regions and regulate bodily movements in relationship to reward processing, aversive stimuli, and the
sleep-wake cycle.
The ‘mindlock’ state is characterized by periods of either global akinesia or bradykinesia of the upper
limbs; conjugate contraversive horizontal eye movements, an absence of vertical eye movements, and
minimal eye blinks. Reduction of facial and skeletal muscle tone and loss of response to aversive
stimuli occurred and prolonged periods of DBS led to a ‘sleep-like’ state with spindle oscillations.
Hyperkinetic stereotyped movements of the upper limbs and ‘tic-like’ behaviors were observed at DBS
offset, lasting several seconds prior to the animals’ resumption of behavioral performance.
”Mindlock’ exhibited a marked enhancement of ‘beta-band’ (15-25Hz) spectral power recorded across
the ECoG array and LFPs recorded within frontal cortex and medial thalamus. Low frequency (1-15Hz)
and higher frequency (25-100Hz) power were significantly attenuated. These findings bear strikingly
similarities to the symptoms and EEG signal abnormalities observed in human subjects suffering from
basilar migraine with exception of immediate cessation upon DBS offset. The robust and reproducible
effects reported here may provide a mechanistic insight into this rare syndrome and in addition, lends
support for further evaluation of this target, specifically the LHb and its proposed role in psychiatric
conditions that includes depression.
Poster Number: NM-103
Poster Name: Simultaneous High-definition Transcranial Direct Current Stimulation and Motor
Imagery Acutely Modulates Activity in the Motor Cortex
Authorship List: Bryan Baxter, Bin He
Institution: University of Minnesota
Abstract: Transcranial direct current stimulation (tDCS), a noninvasive electrical neuromodulation
technology, has been found to improve explicit and implicit motor learning. Sensorimotor Rhythm
(SMR) based Brain-Computer Interfaces (BCIs) utilize a subject’s imagined movement to control virtual
and physical objects. A similar sensorimotor network is active during both motor movement and motor
imagery. With this parallel between motor movement and imagery, we hypothesize that anodal tDCS
can be used to improve the learning of motor imagery based BCIs. Two sets of experiments were
performed: anodal/cathodal/sham stimulation using a 4x1 HD-tDCS combination over the left
sensorimotor cortex during motor imagery performance without feedback and anodal/cathodal/sham
stimulation with feedback during BCI performance. For the within subject motor imagery experiment,
we found a significant interaction effect in electrode C3, over the left sensorimotor cortex, eventrelated activity between stimulation type and the time block (prestim, during stim, immediately poststim, and 30 minutes post-stim) within a single session of each type of stimulation. For the BCI
experiment within a single session, we found a significant interaction effect of event-related activity
between anode and sham stimulation during BCI learning within a single session in electrode C3, but
not C4. In addition, over the course of 3 sessions of BCI learning, subjects who underwent anodal
stimulation tended to have higher accuracy compared to those who had the sham stimulation. These
results illustrate that within a single session, tDCS stimulation alters the neural activity underlying
motor imagery and may improve motor imagery learning.
Poster Number: NM-104
Poster Name: Hybrid Automated Computer Program for Localization of STN Borders during DBS
Surgery
Authorship List: Kevin M. Biddell, George T. Mandybur
Institution: University of Cincinnati, Mayfield Clinic, Biosynesi LLC
Abstract: Accurate detection of the STN is paramount in STN DBS surgery. Presently subjective data
is used to determine STN border and cellular activity. This may impart inconsistencies in the
procedure. Automated detection can standardize this process. The objective of this project was to
develop a computer program to compare functional algorithms that automatically recognize the
boundaries of the STN by analyzing microelectrode recordings (MER) made during DBS lead
placement. We believe that a hybrid of multiple methods can be used to develop a more sophisticated
automatic localization tool. Our team has developed a Matlab program to be used as a test bed that
simultaneously runs multiple automatic localization algorithms. The program is adaptable and
expandable. We present retrospective data on 13 patients with a total of 28 tracts who underwent
MER guided STN DBS implantation. The MER data was analyzed by two automatic methods based on
the Root Mean Square (RMS) and the Power Spectral Density (PSD), both of which have been
compared with the physician’s intraoperative determination of the STN borders. Multiple data
enhancement methods were incorporated in the program and used in the analysis presented. Both
methods detected STN dorsal and ventral boundaries earlier then the neurologist. The RMS method
estimated STN entry at 0.429 mm and exit 1.33 mm on average earlier then the physician, and the PSD
method calculated entry 0.703 mm earlier and the exit 1.35 mm earlier, on average. With these
methods we will discuss the implications and possible standardization of STN border detection.
Poster Number: NM-105
Poster Name: Directly Implantable Biocompatible Liquid Crystal Polymer Structures for
Neuromodulation Applications
Authorship List: E. Bihler, D. Schulze, R. Toelke, S. Bagen
Institution: Dyconex AG, Micro Systems Technologies, Inc.
Abstract: Liquid crystal polymer (LCP) dielectric can be used to form stable biocompatible structures
for medical implants without the need for hermetic coatings or housings.[1,2] This is particularly useful
for producing lead structures for neuromodulation applications where miniaturization is critical to
successfully stimulating specific brain and spinal cord neurons or peripheral ganglia. Accompanying
conductor materials are chosen based on the intended duration of the implant. For implants of < 30
days, inner conductor layers of copper are acceptable provided they are adequately encased within
the near-hermetic LCP and any exposed conductor surface is capped in a noble metal such as gold.
For implants of > 30 days, only noble metals such as gold, titanium and platinum are acceptable to
ensure long-term biocompatibility. In addition, processing chemicals and operations used to fabricate
the implant structure must not induce toxic contamination.[3]
References
[1] J. Jeung, et al., "A novel multilayered planar coil based on biocompatible liquid crystal polymer for
chronic pain implantation," Sensors and Actuators A: Physical, Volume 197, 1 August 2013, pp. 38-46.
[2] S.W. Lee, et al., "Development of Microelectrode Arrays for Artificial Retinal Implants Using Liquid
Crystal Polymers," IOVS, December 2009, Vol. 50, No. 12, pp. 5859-5866.
[3] ISO 10993-1:2009 Biological evaluation of medical devices – Part 1: Evaluation and testing with a risk
management process.
Poster Number: NM-106
Poster Name: Evaluation of Interhemispheric Effective Connectivity in Chronic Stroke using TMSEEG
Authorship List: M. R. Borich, L. A. Wheaton, S. M. Brodie, B. Lakhani, L. A. Boyd
Institution: Emory University, Georgia Institute of Technology, University of British Columbia (Canada)
Abstract:
Introduction: Reorganization and remodeling of motor network connections contributes to recovery of
arm function after stroke. Changes in effective network connectivity in humans after stroke can be
studied using concurrent transcranial magnetic stimulation (TMS)-electroencephalography (EEG). The
primary objective of this study was to use imaginary coherence (IC) analysis of TMS-evoked EEG
responses to directly characterize interhemispheric M1 effective connectivity in individuals with stroke.
Methods: Ten participants with chronic ischemic stroke in the right (n=5) or left (n=5) hemisphere and
four matched controls were tested. Standard TMS procedures were conducted bilaterally.
Transcallosal inhibition (TCI) was evaluated using single TMS pulses over ipsilesional M1 while
performing an ipsilateral grip force contraction (50% maximum). 64-channel EEG recordings were
collected during TCI. TCI was evaluated from the non-dominant hemisphere in matched controls.
All standard data pre-processing steps were performed in EEGLAB. Epoch were extracted for each
participant and concatenated within each group for IC analysis. Post-TMS IC values between
electrodes overlying M1 bilaterally were calculated with the beta frequency band (15-30Hz) as the
primary dependent measure of interhemispheric IC.
Results: Individuals with chronic stroke showed greater TMS-evoked EEG beta IC (p<0.05).
Participants with lesions in the left hemisphere (n=5) exhibited greater arm impairment compared to
individuals with right hemisphere lesions (n=5). Greater interhemispheric beta IC was observed in
those with left hemisphere lesions.
Discussion: Preliminary findings suggest there is increased interhemispheric effective M1 connectivity
during an active motor state that may be behaviorally relevant to levels of arm impairment in chronic
stroke.
Poster Number: NM-107
Poster Name: High Density Penetrating Electrodes for Autonomic Nerves
Authorship List: J. Burns, A. Mueller, D. Chew, A. Achyuta, J. Fiering
Institution: Charles Stark Draper Laboratory
Abstract: Electrode arrays for recording and stimulation in the central nervous system have enabled
numerous advances in basic science and emerging therapies. In particular, microfabricated arrays with
precision relative positions and electrode size offer the benefit of accessing single neurons and permit
precise mapping of neuron function. Similar advances can be envisioned toward understanding the
autonomic nervous system and developing therapies based on its modulation, but appropriate
electrode arrays are currently lacking. Here, we present for the first time a multichannel electrode
array suitable for penetration of peripheral nerves having diameters as small as 0.1mm. The arrays offer
the potential to access multiple discrete nerve fibers within such small nerves. We fabricated and
characterized 5-channel arrays and obtained preliminary recordings of activity when penetrating rat
vagus nerve. The electrodes were constructed using hybrid microfabrication processes, and three
variations of the approach were compared, each using a different electrode metal. The individual
electrode shafts are as small as 0.01mm in diameter and at its tip each has a defined site that is
addressable via a standard electronic connector. In addition to acute in vivo results, we present
evaluation in saline, including electrochemical impedance spectroscopy. Having established the
fabrication method, our next steps are to incorporate the arrays into an implantable configuration for
chronic studies, and here we further describe concepts for such a device.
Poster Number: NM-108
Poster Name: Neuroimaging Biomarkers of Relapse in Addiction — Potential Neuromodulation
Intervention Targets
Authorship List: Jazmin Camchong, Matt Kushner, Teresa Kimberley, Kelvin O. Lim
Institution: University of Minnesota
Abstract: The relapsing nature of alcoholism is a major obstacle to successful treatment. New
interventions targeting brain biomarkers of relapse hold significant promise in reducing this critical
public health problem. Data suggest that nucleus accumbens (NAcc) functional connectivity (FC),
particularly low NAcc-prefrontal FC, is a potential brain biomarker of relapse vulnerability. To
investigate whether NAcc-PFC FC can be modulated, we will conduct a double-blind placebocontrolled randomized study using transcranial direct current stimulation (tDCS; active and sham
conditions) on 20 individuals with alcohol use disorder (AUD; 18-45 years old; 3 weeks abstinent) from
the Lodging Plus Treatment Program. AUD will be randomly allocated to receive either (i) active tDCS
or (ii) sham tDCS while performing a reversal learning task, twice a day (separated by at least 3 hours)
on 5 consecutive weekdays. We will compare (i) magnitude of change in NAcc-PFC FC and (ii) change
in craving scores reported before and after intervention between groups. We will explore whether
changes in craving (between 2 and 3 weeks of abstinence) are a potential predictor of subsequent
relapse (6 months later). To correlate neuromodulation intervention with long-term clinical outcome,
we will record craving and relapse status during the 6 months following treatment discharge. We will
examine the relationship between change in NAcc-PFC FC between 2 and 3 weeks of abstinence and
subsequent (i) monthly craving scores and (ii) relapse status during follow-up. We hypothesize that
PFC stimulation will (i) enhance FC between NAcc and PFC and (ii) reduce craving in addiction.
Poster Number: NM-109
Poster Name: EEG Informed fMRI Analysis in Sickle Cell Disease Patients during Resting State
Authorship List: Michelle Case, Clara Huishi Zhang, Yvonne Datta, Stephen Nelson, Kalpna Gupta, Bin
He
Institution: University of Minnesota
Abstract: Sickle cell disease (SCD) is the most common inherited blood disorder in the US. Pain is the
most common reason for SCD patients to be hospitalized. However, pain treatment remains suboptimal partly due to inadequacy of pain measurement. Therefore, a quantitative objective imaging
method to measure pain is needed to help treat SCD and other chronic pain patients. The goal of this
study is to find biomarkers of chronic pain. Eight SCD patients and five healthy controls were
recruited. All subjects have given written consent, and all procedures have been approved by the IRB
of the University of Minnesota. Simultaneous recordings of functional magnetic resonance imaging
(fMRI) and electroencephalography (EEG) during resting state for SCD patients and healthy controls
were obtained. EEG was used to inform fMRI results. Spontaneous power fluctuations over different
frequencies ranging from 1-30 Hz and EEG microstates were calculated. The time courses were used
to obtain t-contrast statistical maps. Frequency power and microstates show that healthy controls
showed the default mode network (DMN) activity, while the SCD patients showed reduced DMN
activity. DMN was negatively correlated with alpha band in controls and reduced DMN was negatively
correlated with beta band for SCD patients. EEG microstates also reflected diminished DMN activity
in SCD patients. This study suggests that DMN activity can be used as a potential pain biomarker.
These EEG methods can be used as features to help classify pain in future studies.
This work was supported in part by NIH U01-HL117664 and NSF IGERT grant DGE-1069104.
Poster Number: NM-110
Poster Name: Repetitive Transcranial Magnetic Stimulation Changes Kinematic Performance in
Children with Perinatal Strokes
Authorship List: Chao-Ying Chen, Karah Bush, Warren Lo, Ajit Chaudhari, Michael McNally, Jill
Heathcock
Institution: University of Minnesota
Abstract:
Background & Purpose: Perinatal stroke can result in neurological impairments that interfere with
children’s abilities to participate in functional tasks. A growing body of evidence describes transcranial
magnetic stimulation (TMS) as a treatment that has had favorable outcomes on upper extremity
function in children with hemiparesis. The purpose of this case series is to describe the change in
upper extremity function following repetitive transcranial magnetic stimulation (rTMS) in children with
motor impairments from a perinatal stroke.
Case Description: Four children (mean age=14.5 year-old) were randomly assigned to undergo either
rTMS (n=3) or sham control (n=1) sessions. Eight rTMS or sham sessions were performed over a twoweek interval. rTMS was administered to inhibit unaffected hemisphere at a frequency of 1 Hz for 20
minutes. Subjects underwent a series of assessments including kinematic performance during a
reaching task, grip strength, and Shriner’s Hospital Upper Extremity Evaluation (SHUEE).
Outcomes: Preliminary kinematic results indicate a trend of greater improvements in symmetry for
movement units between involved and non-involved side in the rTMS group. Increased grip strength
and SHUEE scores were observed only in the rTMS group. Grip strength for the rTMS group increased
by a range of 18-167% (1.34-3.34 kg) in the involved limb. SHUEE scores, specifically for the wrist and
elbow segments, increased by a range of 10-25% and 0-50% respectively following rTMS.
Conclusion: Preliminary findings indicate that rTMS may improve reaching quality, hand strength and
mobility in children with hemiparesis.
Poster Number: NM-111
Poster Name: The Effects of Caffeinated and Sugared Soft Drinks on Youth Brain
Authorship List: John Chen, Jeffrey He, Albert You
Institution: Wayzata High School, Moundsview High School, University of Minnesota
Abstract: Caffeine and sugar are two common ingredients used for most soft beverages. Despite some
concerns about potential side effects linked to over consumption of caffeinated or sugared soft drinks,
Americans consume more than ever before, especially teens and young adults. One possible
explanation to this accelerated upward trend of soft drink consumption is the lack of solid scientific
evidence that shows the effects of soft beverages on our body, especially on our brain. What we want
to find out in this study is how the intake of caffeine contained in a standard sized can (12 ounce) of
Diet Coke or the sugar contained in a 12-ounce Sprite, two of the most popular soft beverages from
the largest soft drink retailer (Coca-Cola) company, actually affects the brain activity of young people
through the use of a 64-channel electroencephalography (EEG) recording facility located at the
University of Minnesota. We found significant effects on the brain alpha-wave EEG activity with
opposite trends in response to the intake of sugar-dominant (upward trend) and caffeine-dominant
(downward trend) soft drinks. The induced brain EEG activity changes were short-lived; it wore off
within one hour. These findings may provide a clue about the tendency for over-drinking pop
beverages if the mindset is not aware about the possible side effects of the large amounts of
accumulated caffeine or sugar contained in the popular beverages, in particular, for young and
immature brains. Therefore, public awareness is important for proper control of daily soft drink
consumption.
Poster Number: NM-112
Poster Name: Reliable Seizure Prediction Using EEG Data
Authorship List: Vladimir Cherkassky, Brandon Veber, Jieun Lee, Ned Patterson, Gregory A. Worrell,
Benjamin H. Brinkmann
Institution: University of Minnesota
Abstract: There is a growing interest in data-analytic modeling for prediction and/or detection of
epileptic seizures from EEG recording of brain activity. Even though there is clear evidence that many
patients have changes in EEG signal prior to seizures, development of robust predictive methods
remains elusive. We argue that the main issue for development of effective EEG-based predictive
models is an apparent disconnect between clinical considerations and data-analytic modeling
assumptions. We present an SVM-based system for seizure prediction, where design choices and
performance metrics are clearly related to clinical objectives and constraints. This system achieves
very accurate prediction of preictal and interictal EEG segments in dogs with naturally occurring
epilepsy. However, our empirical results suggest that good prediction performance may be possible
only if the training data set has sufficiently many preictal segments, i.e. at least 6 seizure episodes.
Poster Number: NM-113
Poster Name: MVPA for Deep Brain Stimulation: Characterization of fMRI Activation Pattern in STN
and GPi DBS
Authorship List: Shinho Cho, Paola Testini, Megan Settell, Erika Ross, William Gibson, Kevin Bennet,
Kendall Lee, Hoon-Ki Min
Institution: University of Minnesota, Mayo Clinic
Abstract: Despite the widespread use and efficacy of Deep Brain Stimulation (DBS) in movement and
neuropsychiatric disorders, its mechanisms of functioning are still not fully understood. Functional MRI
could provide the appropriate temporal and spatial resolutions to investigate the neural modulation
effect evoked by DBS. Here we showed how individualized DBS-evoked Functional Brain Map (FBM)
could be developed by using multi-voxel pattern analysis (MVPA) which has emerged as a useful
analysis to infer spatiotemporal neural activity.
We performed subthalamic nucleus or globus pallidus internus DBS in a large swine population [STN
(n=7) vs GPi (n=7)], and then classified multiple regions of voxel activation patterns according to each
stimulation category, which are defined by electrode contact information. DBS-evoked FBM based on
the selected ROI voxels was generated and fed to soft margin Support Vector Machine (Cortes &
Vapnik, 1995) for training and test.
We found the SVM model could accurately classify DBS-locus in significantly higher than the chance
level based on the combination of 42-ROI activation patterns. The classification accuracy was about
80% in average after the cross-validation procedure. More interestingly, the accuracy was still
maintained over 70% across different swine group (cross-subject validation), indicating that the model
could be generalized to further new subject group.
In conclusion, our results support the efficacy of MVPA in developing DBS-evoked FBMs. For further
researches, clinical outcomes of individual human subjects associated with specific targeting could be
predicted based on the comparison of the associated DBS-evoked FBM with a previously established
FBM database.
Poster Number: NM-114
Poster Name: Joint Brain Connectivity Estimation from Diffusion and Functional MRI Using a Network
Model
Authorship List: Shu-Hsien Chu, Christophe Lenglet, Keshab K. Parhi
Institution: University of Minnesota
Abstract: A brain communication network model is presented in this paper to benefit the anatomical
connectivity estimation from joint diffusion MRI (dMRI) and functional MRI (fMRI) analysis. More
specifically, the goal is to improve the estimation of anatomical connectivity patterns by the
constraints derived from functional information. fMRI measures the synchronizations of brain activities
and the synchronizations are the result from message exchange that is carried on the axon fiber
bundles, i.e., anatomical connections. Moreover, the connection strength can be estimated using
conventional tractography techniques. Therefore, in the network, the grey matter parcels are
considered as the nodes and the axon bundles are analogized to the edges. The relations between
activity coherence, message exchange, and fiber capability are formulated in the information
requirement constraint, sharing constraint, and capacity constraint. Solving the network optimization
problem finds the connections that are missed or weakened in the tractography recovery and extract
the anatomical pathways, brain micro structure, for the functional connections, brain macro structure.
The validation and comparison to the prior work on a realistic phantom show that the proposed
network model has the best correct rate with the least false-positive rate. The brain result is
demonstrated using Human Connectome Project data. In the brain result, the proposed model
consistently finds many U-shape curved connections and inter-hemisphere connections that are either
relatively weak or not even shown the result of conventional tractography technique.
Poster Number: NM-115
Poster Name: Mining Performance-related Neural Co-modulators during Driving by Applying
Independent Co-modulation Analysis and Fuzzy Neural Network
Authorship List: Chi-Yuan Hsieh, Shang-Wen Chuang, Chun-Hsiang Chuang, Chin-Teng Lin
Institution: National Chiao-Tung University (Hsinchu, Taiwan)
Abstract: Recently, independent component analysis (ICA), a widely used tool for biomedical signal
processing, was applied to EEG data to identify temporally distinct (independent) signals generated by
partial synchronization of local field potentials within cortical patches. In the previous study, spectral
activities of many maximally “temporally” independent brain processes were observed to covary with
the fluctuation of task performance, suggesting the existence of the co-modulatory brain dynamics.
This research applied the 2nd independent component analysis on spectra of the EEG cortical sources
to decompose the performance-related neural co-modulators during a driving task. We also applied a
non-linear method - Self-Constructing Neural Fuzzy Inference Network (SONFIN) to assess the
relationship between task-related independent modulators and task performance. SONFIN is a
general connectionist model of a fuzzy logic system, and has ability to compile an optimal structure and
tune parameters automatically by an on-line learning mechanism without requiring any assignment of
fuzzy rules in advance. Results showed that the correlation coefficients of the Delta-Theta
independent modulator (IM) were higher than those of Delta IM, Alpha IM and Beta IM. The evidence
provided in this study further demonstrated the neural mechanism of driving performance.
Poster Number: NM-116
Poster Name: Multichannel Transcranial Magnetic Stimulation: A Simulation Study
Authorship List: Christopher C. Cline, Bin He
Institution: University of Minnesota
Abstract: Transcranial magnetic stimulation (TMS) is a neuromodulation technology that allows for
noninvasive focal stimulation of the human cortex. During the technology’s early development in the
1980s, TMS devices initially used a single circular coil for stimulation, but quickly advanced to using
figure-8 coils with significantly improved focality. 35 years later, the figure-8 coil is still the standard in
research and clinical use. Many other coil designs have been explored for optimizing focality, but few
have demonstrated a significant benefit over the conventional figure-8 design. These designs largely
utilize a single-channel stimulator to provide current to all coil windings. However, adding independent
control of multiple coils using a multichannel stimulator can greatly improve the flexibility of a TMS
system. We propose a multichannel TMS (mTMS) system that makes use of subject-specific anatomical
information for improved targeting of stimulation. This work uses finite-element modeling with
individualized head models to predict the electric fields induced in the cortex by an mTMS coil array,
and demonstrates the use of numerical optimization of individual coil currents to obtain desired
stimulation profiles. By taking into account subject-specific cortical geometry, this approach can
produce significantly more focal stimulation profiles than a standard figure-8 coil. Additionally, the site
of stimulation in a particular subject can be adjusted by redistributing driving currents between coils,
without physically moving the TMS coil array.
Poster Number: NM-117
Poster Name: Burst and Tonic Spinal Cord Stimulation Attenuate Pain via Differential Mechanisms in a
Rat Model of Cervical Radiculopathy
Authorship List: N. D. Crosby, C. L. Weisshaar, J. R. Smith, M. E. Zeeman, M. Goodman Keiser, B. A.
Winkelstein
Institution: University of Pennsylvania
Abstract: Because preclinical studies report some differences between burst and tonic spinal cord
stimulation (SCS), it is possible that these two modes of SCS work through different mechanisms.
Since tonic SCS has been shown to activate inhibitory GABA signaling in the dorsal horn, we compared
the effectiveness of burst and tonic SCS in the context of GABA signaling, using a rodent model of
radiculopathy. The effects of burst and tonic SCS were compared by recording dorsal horn neuronal
excitability before and after each mode of stimulation at day 7 following a painful cervical root
compression in Holtzman rats (n=8). Neuronal firing was also recorded before and after each SCS in
the presence of the GABAB receptor antagonist, CGP35348 (n=7). In a separate study, spinal cord
stimulators were implanted to deliver either burst (n=5) or tonic (n=7) SCS beginning on day 4 after the
painful root compression and remained on until day 14. Burst and tonic SCS both reduced spinal
neuronal firing at the 90% motor threshold. The effect of tonic SCS (p<0.05), but not burst SCS, on
neuronal firing was blocked by CGP35348. Although both modes of SCS attenuated pain, only tonic
SCS attenuated injury-induced decreases in serum GABA, maintaining levels at baseline. In contrast,
serum GABA levels remained decreased from baseline during burst SCS (p<0.02), similar to levels
without SCS. These studies suggest that burst SCS does not act via spinal GABAergic mechanisms,
despite observed similarities in the attenuation of spinal hyperexcitability and pain by both burst and
tonic SCS.
Poster Number: NM-118
Poster Name: Minnesota Magnetic Brain Array for Stimulation and Mapping of Brain Cells
Authorship List: Mahendra DC, Ang Klem, Yinglong Feng, Mahdi Jamali, Susan Keirstead, Walt Low,
Jian-Ping Wang
Institution: University of Minnesota
Abstract: Minnesota Magnetic Brain Array (MMBA) is a novel device to modulate/stimulate neurons
and other biological components which are sensitive to magnetic fields. The proposed design for
MMBA and theoretical verification of operation presented here are critical steps in being able to map
3D brain structures and overcome neurological disorders such as Parkinson, dystonia, depression etc.
MMBA consist of novel magnetic nanostructures combined with magnetic sensors. Magnetic
nanostructures produce fluctuating magnetic fields which are used to induce axial current in the
neuron for the stimulation of nerve cell. MMBA can locally activate neurons and offers benefits such
as low power consumption and high density compared to state-of-art electrode based neuron
modulation arrays. We calculated the axial current in the axon by using the core-conductor model and
show current flow in a single neuron. To induce this amount of current for the stimulation of a nerve
cell the amount of magnetic field required is 0.418 pT. Our analytical calculation and micromagnetic
simulation showed that the field generated by a novel magnetic nanostructure we proposed (of the
order of 0.1 mT) can be used to induce axial current for the stimulation of nerve cells.
Poster Number: NM-119
Poster Name: Transcranial Direct Current Stimulation with a Bi-lateral Ring Electrode Enhances
Recovery of Motor Performance following Neonatal Hypoxic-Ischemic Stroke in Rats
Authorship List: Thomas DeMarse, Christopher L. Anderson, Martha Douglas-Escobar, Candace
Rossignol, Adityakumar Kasinadhuni, Rachel Nelson, Thomas Mareci, Rosalind Sadleir, Michael King,
Prodip Bose, Michael Weiss, Paul Carney
Institution: University of Florida
Abstract: Perinatal hypoxic-ischemic encephalopathy (HIE) stroke affects 1 infant for every 1600
children born, is a leading cause of death in infants, and results in severe impairments including
epilepsy, developmental delay, and cognitive-motor deficits. In this study we tested the therapeutic
effect of daily repeated non-invasive tDCS stimulation episodes using a novel bi-lateral ring electrode
upon the motor performance of infant (P7) rat HIE stroke models (Rice-Vannucci). Forty three rat pups
were separated on P21 into three groups: tDCS treated (n=24), Non-stimulated (n=19), and Sham “nonstroked” controls (n=12). A ring electrode was fixed to the skulls of animals in each group and
stimulated with anodic current daily (13 on/3 off/13 min on at 200 μA with 3 min ramp up and down)
under isoflurane for one week. Neurological & Motor performance assays were performed including
negative geotaxis, grip strength, and gait analysis were assessed each day. Following stimulation and
behavioral tests, tissue and protein samples were collected and used to measure brain-derived
neurotropic factor (BDNF) levels in a small sample group of stimulated and non-stimulated animals.
Motor performance in tDCS treated HIE animals improved significantly over the course of one week,
was correlated with a return to near-normal body weight, strength, gait, and motor function compared
to the non-treated HIE controls, and was associated with large-scale increases in BDNF expression at
P30 and P36 reinforce the role of plasticity for therapeutic effect.
Poster Number: NM-120
Poster Name: Event-related Desynchronization of Independent Motor-related EEG Components in
Different Motor Tasks
Authorship List: Jeng-Ren Duann, Jin-Chern Chiou
Institution: China Medical University (Taichung, Taiwan)
Abstract: This study compared the robustness of the event-related desynchronization (ERD) induced
by three motor conditions, namely motor execution (ME), motor imagery (MI) and motor observation
(MO). During a 10-min EEG experiment, 13 healthy right-handed subjects were asked to perform 60
trials of three motor tasks (20 trials for each condition). The sequence of the motor tasks were
randomly assigned. During the ME condition, the subjects were asked to physically make a fist with
their left hand once during a 3-sec window. For other conditions, they would either imagine making a
fist (MI) or passively watch a video clip of someone making a fist (MO), both with left hand. EEG data
were acquired using a NeuroScan 64-channel SynAmp2 system, at the sampling rate of 1000 Hz (bandpass filtered between 0.1 - 250 Hz). The EEG data were first decomposed into independent EEG
components using independent component analysis (ICA), such that we could look at pure motorrelated processes with minimal contamination. The component associated with the motor-related
activity was selected for each subject. From the group mean ERD plot, we selected the peak ERD
frequency band (alpha, 8-13 Hz) to compute the conditional ERD and compare the resultant ERD
across different motor condition. The result showed that ME exhibited the deepest (~-8 dB) and
longest suppression (~ 2 sec) than the other two. MI created only brief suppression with half the
amplitude (~-4 dB), which was slightly more reliable than the suppression induced by MO.
Poster Number: NM-121
Poster Name: Validation of 7T MRI Patient-specific DBS Lead Localization With MER
Authorship List: Yuval Duchin, Guillermo Sapiro, Kenneth Baker, Jon McIver, Jerrold Vitek, Noam
Harel
Institution: University of Minnesota
Abstract:
Background: Structural images acquired at 7 Tesla (T) exhibit rich informational content with potential
utility for clinical applications. Here we utilized 7T images to create patient-specific anatomical models
to enhance pre-surgical DBS targeting as well as post-surgical visualization of the DBS lead location.
Methods: Nineteen candidates for DBS surgery were scanned preoperatively on 7T MRI systems and
3D volume rendering of the anatomical target were generated. Following standard clinical procedure,
serial microelectrode recording (MER) techniques were used intra-surgically to map the target region
and optimize DBS lead placement. A postoperative CT was co-registered to the preoperative 7T MRI.
The registered images were then used to assess electrode and intraoperative microelectrode locations
relative to the 3D anatomical model and compared with MER mapping and clinical programming data.
Results: The data indicates excellent agreement between the 7T anatomical model, the intra-operative
MER mapping and the post-operative electrode programing configuration. A pooled collection of the
activated contacts in individual patients was computed across all patients and a cluster of active
contacts was identified at the dorsolateral aspect of the STN.
Conclusion: Structural 7T MRI can be used to create accurate, patient-specific models for use in DBS
procedures. Specifically, these models may be of use for (1) direct targeting; (2) verify final DBS lead
and contacts location; 3) facilitate DBS programming for maximum benefit to the patient; and (4) allow
for further understanding of the optimal location of the DBS electrode within the target region.
Poster Number: NM-122
Poster Name: Using EEG Source Imaging to Expose Self-modulated Rhythmic Activity during Right
Hand Motor Imagery Tasks
Authorship List: Bradley Edelman, Bryan Baxter, Taylour Hanson, Bin He
Institution: University of Minnesota
Abstract: Control of real and virtual devices using sensorimotor-based (SMR) brain-computer
interfaces (BCIs) has achieved successful and high dimensional control. SMR BCI control signals are
founded on the user’s ability to increase or decrease frequency-specific neuronal activity in the
primary motor cortex caused by the imagination of motor movement. These phenomena can be
successfully separated; however, the imaginations used to generate them often have little to do with
the output command, revealing a cognitive disconnection between the user’s motor intent and the end
effector’s action. Therefore, there is a need to develop techniques which can identify with high spatial
resolution the self-modulated neuronal activity reflective of the actions of a helpful output device.
Over the past decade EEG source imaging (ESI) techniques have proven to be an effective approach
for interpreting motor intent by reconstructing the current density on the cortical surface. We extend
previous ESI work to natural hand/wrist manipulations by applying a novel technique to classifying four
complex motor imaginations of the right hand, flexion, extension, supination and pronation. It is our
hypothesis that ESI techniques can increase the spatial resolution of the scalp to expose and exploit
smaller regions of varying electrical activity to better classify functional motor imagery tasks of the
right hand. We report an increase of up to 12.2% for individual task classification and 8% for overall
classification using the proposed ESI approach over the traditional sensor-based methods. The
successful development of a natural BCI system may help subjects for self-modulation of brain states.
Poster Number: NM-123
Poster Name: PDMS-based Optical Conduit for Powering Untethered Neural Microstimulators
Authorship List: Ali Ersen, Mesut Sahin
Institution: New Jersey Institute of Technology
Abstract: Emerging applications in neural prosthetics such as chronically implanted microstimulators
demand flexible, biocompatible, untethered devices. Light activated micro-electrical stimulators have a
potential to improve the longevity of implants, reduce the device size, and minimize the tissue
response by eliminating the tethers. These stimulators require delivery of optical energy to the site of
implant, which may be several centimeters away from the skin or an extracorporeal light source.
Polydimethylsiloxane (PDMS) has excellent mechanical and optical properties as a material for making
implantable optical fibers or conduits. This study describes the design, fabrication and testing of a
PDMS based light conduit to optically energize untethered micro-electrical stimulators for activation
of neural tissue. This optical conduit can deliver light into depths of neural tissue within a wide range of
wavelengths including those used in optogenetics. The softness and flexibility of PDMS as compared
to harder materials used in optical fibers is expected to induce minimal trauma to the surrounding
tissue.
The core and cladding were fabricated using PDMS with difference refractory indices; one with a
relatively higher index for the core with RI=1.55 (Dow Corning OE-6550) and another with RI=1.41 for
cladding (Dow Corning Slygard 184). The core was cured for 8 hours at 60°C and the cladding was
cured for 2 hours at 80°C.
For preliminary testing, a 100mm long core with a rectangular cross-sectional area of 2000µm x 1500µm
at one end and tapering down to a size of 500µm x 500µm at the other end was fabricated. The
cladding had a thickness of about 250µm. The conduit was coupled to a red laser diode (Flexpoint Mini
Laser, 650nm, 4.9mW) and the light exiting from the smaller end was collected with a photodiode
(Thorlabs, Silicon Photodiode, 10mm x 10 mm active area) for measurements of attenuation at various
lengths by cutting 5mm of the conduit each time.
The preliminary results demonstrated that the conduit has a good confinement of red light and
relatively low attenuation of approximately 0.4-0.5 dB/cm in straight positioning. Attenuation under
bent conditions will be studied to evaluate transmission efficiency for cases where the conduit will be
making curves in order to reach deep structures in the CNS.
This work was funded by National Institute of Health/ NIBIB (R01 EB009100).
Poster Number: NM-124
Poster Name: The Modulation of Conditioned Fear Generalization with D-Cycloserine: An fMRI Study
Authorship List: T. Espensen-Sturges, A. Kielbasa, P. Burton, K.Cullen, S. Lissek
Institution: University of Minnesota
Abstract: During conditioned fear generalization, learned fear responses extend to a range of safe
stimuli resembling the conditioned danger cue. Responses create gradients where fear is high for the
conditioned danger cue and rapidly drops off as stimuli dissimilarity increases. Although some amount
of generalization is likely adaptive, overgeneralizing fear to safe stimuli is thought to be a marker of
anxiety disorders. As such, fear generalization may be a potential target for treatments of these
disorders. We used D-cycloserine, a partial agonist of the NMDA receptor glycine site shown to
reduce levels of conditioned generalization in animals, to modulate fear generalization in healthy
participants. Participants received DCS (250 or 500 mg) or a placebo before undergoing acquisition of
conditioned fear involving the presentation of a specific sized ring (CS+) paired with electric shock (US)
and a different sized ring as a safety cue (CS-). Twenty-four hours later generalization was assessed
using the CS+, CS-, and three intermediately sized rings in order to identify effects of DCS-enhanced
memory consolidation on generalization. Due to higher activations to the CS+, brain activation based
generalization gradients were steeper for the 500 mg drug group in the bilateral anterior insula,
supplementary motor area, and caudate. This aligns with behavioral evidence indicating that DCS may
increase fear to conditioned stimuli, suggesting that enhanced memory consolidation of conditionedfear by DCS strengthens the retention of CS+/US association, and renders the memory of the CS+
more readily retrievable in the presence of stimuli that approximate the danger cue.
Poster Number: NM-125
Poster Name: Effect of Different Interstimulus Intervals of Inhibitory Paired Associative Stimulation in
People with Chronic Stroke
Authorship List: Kate Frost, Mo Chen, Jessica Cassidy, LeAnn Snow, Ann Van de Winckel, Teresa
Kimberley, James Carey
Institution: University of Minnesota
Abstract: Noninvasive brain stimulation shows potential for improving post-stroke motor recovery
through neuromodulation of cortical excitability. Paired associative stimulation (PAS), an emerging
method of noninvasive brain stimulation, induces a long-lasting change in neuronal plasticity by pairing
a peripheral nerve stimulus with a cortical stimulus. Due to the role of exaggerated interhemispheric
inhibition in the suppression of cortical activity within the stroke hemisphere, finding the most
effective method of inhibitory PAS applied to the nonstroke hemisphere is desirable. In an effort to
determine an optimal interstimulus interval (ISI) and also to evaluate safety, we applied inhibitory PAS
at three different ISIs (latency-3, -5, or -7 ms) targeting the abductor pollicis brevis of the nonparetic
hand in three individuals with chronic stroke. A fourth condition (latency+100 ms) was used as a sham
control. Change in corticospinal excitability was assessed by comparing the average of 30 motor
evoked potentials elicited by single pulse transcranial magnetic stimulation (TMS) before PAS
treatment and 0, 10, 20, 30, 40, 50 and 60 minutes following treatment. Demonstration of safety was
measured through physician assessment, questionnaire, memory test, as well as nonparetic finger
tracking. Our results showed that an optimal ISI to induce inhibition in the nonstroke hemisphere
exists but may differ between individuals. Also, there were no adverse events and so further
investigations on a larger number of subjects with stroke are now warranted.
Poster Number: NM-126
Poster Name: Ventral Striatum/Ventral Capsule Deep Brain Stimulation Differentially Regulates Mood
and Thalamocortical Connectivity: A Human fMRI Study
Authorship List: William Gibson, Osama Abulseoud, Krzysztof Gorny, Joel Felmlee, Kirk Welker,
Bryan Klassen, Hoon-Ki Min, Kendall Lee
Institution: Mayo Clinic
Abstract: Ventral capsule/ventral striatum (VC/VS) deep brain stimulation (DBS) is an FDA-approved
therapy for treatment-resistant Obsessive Compulsive Disorder (OCD). Acute, stimulation-evoked
increases in mood are often reported by patients receiving the treatment. However, the
neurobiological underpinnings of this intriguing phenomenon remain largely unknown. Patients (n=4)
underwent image-guided stereotactic implantation of bilateral VC/VS DBS leads. Acutely moodelevating DBS parameters were identified in each patient during intraoperative trial stimulation.
Intraoperative fMRI was then acquired during DBS that had positive (2 runs), as well as negative (2
runs) effects on overall mood. Dynamic Casual Modeling (DCM) was employed to examine DBSevoked alterations in effective connectivity between mediodorsal thalamus (MD) and medial prefrontal
cortex (mPFC). In all four patients, both mood-positive and mood-negative VC/VS DBS resulted in
BOLD activation (conjunction, global null; pFWE < 0.05) of MD, lateral orbitofrontal cortex (lOFC),
hippocampus, amygdala, and contralateral cerebellum. However, while mood-negative DBS caused
BOLD activation in mPFC, mood-positive DBS resulted in negative BOLD signal in mPFC, as well as
motor and premotor cortices. Dynamic causal models based on established connectivity of the limbic
corticostriatal loop connecting mPFC and MD were constructed and compared using Bayesian model
selection. The most likely model for mood-negative settings predicted DBS to modulate information
flow from cortex to thalamus via the ventral striatum, while also enhancing the effect of inhibitory
circuits within MD. In contrast, mood-positive DBS was predicted to also modulate extra-striatal
projections from mPFC to MD, while still affecting within-MD connectivity. This study indicates that
VC/VS DBS is capable of exerting parameter-dependent effects on functional and effective
connectivity within circuits strongly implicated in OCD pathophysiology. In addition, the results
suggest a critical role for the mediodorsal thalamic nucleus in mood elevation.
Poster Number: NM-127
Poster Name: Enhancing Cognitive Control using Transcranial Direct Current Stimulation (tDCS)
Authorship List: Casey S. Gilmore, Carolyn L. Gentz, Doris M. Clancy, Molly R. Gierke, Patricia J.
Dickmann, Greg J. Lamberty, Michael T. Armstrong, Kelvin O. Lim
Institution: Defense and Veterans Brain Injury Center
Abstract: Impulsivity is a multidimensional construct that includes a lack of conscientiousness, lack of
premeditation, sensation-seeking, and impaired cognitive control. Impulsivity is observed in a variety of
psychiatric disorders, and typically manifests as impatience, aggression, poor decision-making, and
excessive risk-taking. Previous studies involving healthy subjects have applied transcranial direct
current stimulation (tDCS) bilaterally over dorsolateral prefrontal cortex (DLPFC) - an area involved
with cognitive control functions - inducing a significant decrease on performance measures of risktaking (e.g. Balloon Analog Risk Task (BART) and Risk Task). The purpose of this study is to expand the
current literature on tDCS and risk-taking by exploring the effects of tDCS across a broad range of
subjects who exhibit clinically-relevant impulsivity.
Subjects complete two tDCS sessions per day for five days with additional one and two month followup sessions. Subjects complete self-report questionnaires of impulsivity (e.g. Barratt Impulsivity Scale),
and pre- and post-intervention behavioral measures of risk-taking (e.g. Delay Discounting Task; Risk
Task). Subjects are randomly assigned to receive either active or sham tDCS during performance of
the BART. At the follow-up sessions, subjects complete the questionnaires and risk-taking tasks.
Preliminary results on two subjects suggest that tDCS can effectively reduce risk-taking propensity as
measured by performance on the BART and Risk Task. Subjects’ risk-taking behavior decreased an
average of 9% in the BART and 12% in the Risk Task when comparing the first and last days of active
tDCS treatment.
This study extends previous research that has only included healthy subjects. This study will measure
the magnitude of change on outcome tasks from pre- to post-intervention, and the stability of these
effects over time. Further, this study could have potential application as a non-invasive clinical
intervention for treating patients with decreased cognitive control.
Poster Number: NM-128
Poster Name: Multimodal Synchronization Therapy (mSync): Developing a Noninvasive Approach for
Treating Tinnitus
Authorship List: Cory D. Gloeckner, Craig D. Markovitz, Benjamin T. Smith, Robert J. Hughes, Hubert
H. Lim
Institution: University of Minnesota
Abstract:
Objectives: We propose a noninvasive approach (mSync) for treating tinnitus. mSync attempts to
activate multimodal brain centers using sensory, motor, cognitive and limbic pathways with precise
timing to modulate auditory neurons driving the tinnitus percept. In our previous work, we combined
broadband noise and somatosensory electrical stimulation in animals that led to facilitation or
suppression of auditory activity depending on body stimulation location. In this study, we investigated
the effect of precise timing between somatosensory and acoustic stimulation on spiking activity within
the primary auditory cortex (A1), a region linked to the tinnitus percept.
Methods: We positioned a 32-site electrode array in the right A1 of ketamine-anesthetized guinea pigs
and compared spontaneous and acoustic-driven activity before and after different mSync paradigms.
Subcutaneous needle electrodes were used to electrically stimulate the right mastoid region, left
pinna, and right pinna, and each location was paired with broadband noise stimulation to the left ear
with varying inter-stimulus delays.
Results: Combined with the results from our previous study, mSync with upper and right body sites
induced greater suppression than facilitation of spiking activity, and vice versa. Changes in interstimulus delay altered the ratio of suppression versus facilitation, and this varied depending on the
stimulation location. These encouraging results suggest that varying mSync parameters can
differentially modulate auditory neurons. Further studies are needed in an awake and tinnitus animal
model to verify the therapeutic effects of mSync.
Funding: MD5M Lions Hearing Foundation, NSF IGERT DGE-1069104, NIH NIDCD R03 DC011589, NIH
NCATS UL1 TR000114.
Poster Number: NM-129
Poster Name: Calibration of MEG-DBS in Phantom Head Model
Authorship List: Amelia Gudex, Edgar Peña, Matthew Johnson
Institution: University of Minnesota
Abstract: Magnetoencephalography (MEG) is a non-invasive functional neuroimaging technique that
has the potential to elucidate mechanisms of Deep Brain Stimulation (DBS) due to its high temporal (1
ms) and spatial (~1 cm) resolution. MEG-DBS stimulation-related artifacts have been evaluated in vivo,
though ongoing neural activity obscures artifact characteristics. Here, we characterized the MEG-DBS
artifact in a phantom head model, segmented and 3D printed from the skull of a Parkinson’s Disease
DBS patient. We placed a DBS lead, wire, and extension cable on the head model in the same
configuration as the patient. We evaluated breathing artifacts due to pacemaker movement by
scanning a healthy subject with the pacemaker placed on the chest. At a representative sensor, the
wires and cables alone increased power between 0-17 Hz (20-55%), and DBS stimuli mainly increased
power at the stimulus frequency (57%). Across all sensors, 48-out-of-248 exhibited more than 20%
power increases at stimulus frequency, most of which (40-out-of-48) were directly above the external
coil, wire, and cable (left lateral and left posterior sensors). No aliasing was observed. Pacemaker
movement increased power between 0.13-0.26 Hz (<55%), though these increases were still smaller
than for neural activity and typical physiological noise (>150% between 1-25 Hz). Overall, this phantom
head model will enable precise characterization of DBS artifacts in the absence of neural activity, thus
providing a pre-human testing platform to consider the role of lead extension cables, subcutaneous
routing and coiling of wires, and different stimulus waveforms and patterns.
Poster Number: NM-130
Poster Name: Ultrasound Modulation in Rat Motor Cortex
Authorship List: Daniel W. Gulick, Tao Li, Bruce C. Towe, Jeffrey A. Kleim
Institution: Arizona State University
Abstract: Ultrasound (US) can stimulate motor responses in small animals, and modulate brain activity
in large animals and humans. It shows potential for noninvasive neuromodulation more focal than TMS
or tDCS. However, the mechanism is unknown and there are unexpected limitations: in rats, US can
evoke movement of hindlimbs but not forelimbs.
This experiment sought to clarify the effect of US on rat motor cortex, by testing US modulation of
electrically-evoked movements. A novel 16-channel epidural cortical stimulation (ECS) array was
implanted. The rat was stimulated under anesthesia, with accelerometers on forelimb and hindlimb.
Some electrical stimuli were preceded by US pulses (200 kHz, 60 W/cm2 SPPA, 300 ms, pulsed at 1
kHz, 50% duty). Responses with vs. without US were compared.
In a short-term effect analysis, neither limb showed a significant difference (9 trials). This may be due
to the 400 ms delay between the start of US and the electrical stimulus. However, there was a
significant long-term (several seconds to minutes) effect of US on both limbs (p < 0.0001).
Some US pulses directly evoked hindlimb responses, if more than 3 seconds between pulses. EMG
latency of US-evoked movement was faster than ECS, and comparable to ICMS layer V stimulation.
It is unknown if US modulation occurs by repeated stimulation (similar to rTMS) or an independent
mechanism. This experiment tested modulation and stimulation at the same parameters and brain
region. Results show US can modulate both forelimb and hindlimb motor cortex, despite only
stimulating movement of hindlimb.
Poster Number: NM-131
Poster Name: Methodological Pipeline for Creating Patient-specific Deep Brain Stimulation
Tractography-activation Models
Authorship List: Kabilar Gunalan, Ashutosh Chaturvedi, Yuval Duchin, Guillermo Sapiro, Noam Harel,
Cameron C. McIntyre
Institution: Case Western Reserve University, University of Minnesota, Duke University
Abstract: Deep brain stimulation (DBS) is an established clinical therapy for advanced Parkinson’s
disease. Multiple lines of evidence show that axons within the vicinity of the active electrode contact
are directly activated by stimulation. However, the specific axon pathways directly influenced by DBS
remain difficult to define, especially on a patient-specific basis. Tractography-activation models
(TAMs) provide a methodological framework for identifying axonal pathways that are directly
stimulated by DBS pulses, by combining DBS electric field modeling with tractography-based multicompartment cable models of axons to explicitly quantify the extent of pathway-specific activation.
Therefore, TAMs enable comparison of the theoretical activity of pathways as the stimulation
parameter settings and electrode locations are changed in a patient, facilitating comparisons between
therapeutic and non-therapeutic clinical outcomes. Here we provide a methodological description of
the TAM creation process and workflow, which uses Python to integrate FSL, Freesurfer, SCIRun, and
NEURON. We demonstrate the power of this technique by presenting an example patient-specific
TAM using high-field (7T) structural and diffusion-weighted MRI data, focused on analyzing DBS of the
hyperdirect and internal capsule pathways in a patient with Parkinson’s disease. Our proof of concept
results suggests that at the clinical thresholds for rigidity control the hyperdirect pathway is
preferentially activated, while the internal capsule fibers of passage are only minimally activated.
Therefore, TAMs have the potential to enhance our ability to define stimulation settings that will
provide therapeutic benefit on a patient-specific basis, thereby improving clinical implementation of
DBS technology.
This work is supported by NIH R01 NS085188. KG has been supported by multiple training grants (T32
GM007250, TL1 TR000441, 5T32 EB004314-13, and P200A100112).
Keywords: Deep brain stimulation, Parkinson’s disease, Magnetic resonance imaging
Theme: Imaging, Modeling, and Biomarker
Poster Number: NM-132
Poster Name: Patient-specific Optimization of Combined Pharmacologic and Deep Brain Stimulation
Therapy for Parkinson’s Disease
Authorship List: Reuben R. Shamir, Trygve Dolber, Angela M. Noecker, Benjamin L. Walter, Cameron
C. McIntyre
Institution: Case Western Reserve University
Abstract:
Background: Deep brain stimulation (DBS) of the subthalamic region is an established therapy for
advanced Parkinson’s disease (PD). However, patients often require time-intensive postoperative
management to balance their coupled stimulation and medication treatments. Given the large and
complex parameter space associated with this task, clinical decision support systems (CDSS) based on
machine learning algorithms could represent useful tools to assist in treatment optimization.
Objective: Develop a proof-of-concept implementation of a CDSS that incorporates patient-specific
details on both drug dosages and stimulation parameter settings.
Methods: Clinical data from 10 patients and 89 post-DBS surgery visits were used to create our
prototype CDSS. The system was designed to provide three key functions: 1) information retrieval; 2)
visualization of treatment outcomes, and; 3) recommendation on expected effective therapy with
machine learning methods.
Results: We found that medication dosages, timing of medication administration, and symptom-specific
preoperative response to levodopa were significantly correlated with postoperative outcomes
(p<0.05). Moreover, our analysis reveals that the medications effect on outcomes is of similar
magnitude to that of the DBS therapy. By combining representation of both the medication and
stimulation therapies, the machine learning algorithms accurately predicted 86% (12/14) of the motor
improvement scores at one year after surgery.
Conclusion(s): CDSS that incorporates medications, their timing, and patient-specific symptoms, in
addition to defining theoretically optimal DBS parameter settings, has potential to improve the clinical
management of PD patients.
Poster Number: NM-133
Poster Name: An Investigation into Effectiveness of Different Coil Designs for Transcranial Magnetic
Stimulation on Mice
Authorship List: Priyam Rastogi, R. L. Hadimani, D. C. Jiles
Institution: Iowa State University
Abstract: Transcranial magnetic stimulation (TMS) is a non-invasive treatment for neurological
disorders using a time varying magnetic field. The electric field generated by the time varying magnetic
field is used to depolarize the neurons which can lead to measurable effects. TMS provides a surgical
free method for the treatment of neurological disorders such as depression, post-traumatic stress
disorder, traumatic brain injury and Parkinson’s disease. Before using TMS on human subjects, it is
appropriate that its effects are verified on animals such as mice. The magnetic field intensity and
stimulated region of the brain can be controlled by the shape, position and current in the coils. There
are few reports on the designs of the coils for mice. In this paper, different types of coils are
developed and compared using an anatomically realistic mouse model derived from MRI images.
Parameters such as focality, depth of the stimulation, electric field strength on the scalp and in the
deep brain regions, are taken into account. Correlation was found between depth of stimulation and
electric field strength on the scalp. Coils were compared with different decay rate of magnetic field
and with different penetration levels. These comparisons will help researchers to determine the most
suitable coil design according to their need. This study should result in improvements in the coil
designs hence in the treatment of specific disorders.
Poster Number: NM-134
Poster Name: Effectiveness of dTMS Treatment in Alleviating Symptoms of Major Depression in a
Naturalistic Clinical Setting using H-Coil Technology
Authorship List: Walter Duffy, Zia Choudhry, Ravi Hadimani, David Jiles, Mohammed Waris, Ryan
Nathan, Waquar Siddiqui, Mahesh Rajamani
Institution: Iowa State University
Abstract:
Background: Deep TMS (dTMS) clinical trials using H-coil technology in Major depressive disorder
(MDD) patients has shown significant outcomes following standardized treatment protocols. However,
in naturalistic settings physicians adapt treatment protocols to individual clinical considerations
achieving better outcomes. These varied therapeutic strategies may assist physicians in designing
improved dTMS treatment paradigms. In the current design, we present a review of response and
remission rates to dTMS in MDD patients.
Methods: 93 patients (24% male) with chronic MDD, ages 18-84 years (mean=44.5), previously failing 08 medication regimens (mean=3.9) received dTMS using H-coil technology. Symptom severity was
assessed using PHQ-9 scores at baseline and every 5 treatments subsequently. Patients received an
acute phase dTMS treatment (20 sessions), after which patients received further maintenance
treatment (10 sessions). dTMS efficacy was evaluated using Repeated Measure ANOVA to contrast
baseline, acute and maintenance paradigms. Response (<10 or >50% from baseline) and remission (<5)
were calculated.
Results: A significant treatment effect (p=0.000) on symptom severity was observed in both
paradigms. Post-acute symptom severity average score decreased 47.9 percent (Means=19.02 vs. 9.05)
with 65 patients responding and 40 remitting. Additionally, post-maintenance average scores
decreased 10 percent more (Means=19.02 vs. 7.57) with 70 patients responding and 44 remitting.
Conclusion: These results suggest that, both acute and maintenance dTMS treatment paradigms are
effective in treatment of MDD symptom severity. However, extending the treatment paradigm from 20
to 30 sessions proves beneficial.
Poster Number: NM-135
Poster Name: Epidural Spinal Cord Stimulation: A Novel Therapy in the Treatment of Restless Leg
Syndrome
Authorship List: Marshall Holland, Oliver Flouty, Teri Thompson, Chandan Reddy
Institution: University of Iowa
Abstract:
Objective: To report a unique finding in two patients whose restless leg syndrome (RLS) symptoms
improved following placement of a spinal cord stimulator for failed back syndrome.
Methods: Two patients diagnosed with failed back syndrome, who also suffered from RLS symptoms,
underwent implantation of a spinal cord stimulator following a successful trial. Both were formally
interviewed before and at 6 weeks following their procedure about their RLS symptoms. The patients
also completed the International Restless Legs Syndrome Severity Scale Questionnaire to objectively
quantify the severity of their symptoms.
Results: Both patients reported subjective symptomatic improvement in their RLS symptoms with
improved sleep quality and quantity, in addition to improvement in their back pain. Moreover, the
patient’s IRLSS scores improved following implantation with an average decline in score of 18.5 points
on a 40-point scale.
Conclusions: Epidural stimulation may be an additional alternative therapy in the effort to relieve
sufferers from their symptoms.
Poster Number: NM-136
Poster Name: Predicting Deep Brain Stimulation Parameters to Optimally Disrupt Oscillations in
Parkinson's Disease
Authorship List: Abbey Holt, Max Shinn, Theoden Netoff
Institution: University of Minnesota
Abstract: Deep brain stimulation (DBS) is used to treat motor signs of patients with medicationrefractory Parkinson’s disease (PD). However, tuning stimulation parameters currently requires time
intensive visits to the clinic where a trial-and-error process is used until maximum therapy is achieved
with minimal side effects. There is a need for a systematic approach to tuning parameters based on
patient physiology. With the development of DBS electrodes that can simultaneously stimulate and
record, a closed-loop approach may be taken for tuning stimulation parameters. It is hypothesized that
emergent oscillations in the basal ganglia network, particularly in the beta range (12-35 Hz) lead to
motor symptoms of PD, and that DBS works, at least in part, by disrupting these oscillations. Here we
show that the phase response curve (PRC) can be used to predict stimulation parameters such as
phase, frequency, and waveform to optimally disrupt oscillatory activity. This has the potential to be
used in a closed-loop approach to DBS. We also test a closed-loop approach to DBS by applying
stimulation at certain phases in a computational model of PD. We tested this hypothesis in a computer
simulation study. DBS leads with 80 electrodes were placed deep in the brain, bilaterally. Three
functional nodes were simulated at thalamus, hippocampus and cortex, respectively. Causal relations
among the three nodes were assigned. Applying source imaging techniques and dynamic causal
modeling to the data, the three sources were localized with couple of mm localization error and the
causal nodes and the underlying network dynamics were identified (the interactions and causal links
between the nodes). After this relation was derived, the desired current density to target these nodes
and consequently interrupt pathological activity was achieved by determining optimal current
distribution of the DBS leads via solving the inverse problem. Target areas were hit with good
specificity. In summary, our simulation results suggest good specificity and resolution in selectively
sensing brain networks and targeting them using high density DBS leads.
Poster Number: NM-137
Poster Name: Feasibility Study of an EEG-based Associative Facilitation Paradigm for Chronic Stroke
Patients
Authorship List: J. Ibáñez, J. I. Serrano, M. D. del Castillo, E. Monge, F. Molina, J. L. Pons
Institution: Spanish National Research Council (CSIC) (Madrid, Spain)
Abstract: We performed a feasibility study to analyse the rehabilitation effects of a novel EEG/FESbased associative facilitation paradigm. Four patients completed eight sessions during one month. In
each session, patients performed self-paced reaching movements with the affected upper-limb and
FES was delivered every time the movement intention state was detected by an EEG-based system.
Metrics regarding functional and neurophysiological changes were evaluated in each patient. Overall,
patients could reliably control the interface by naturally performing the movements (80.9%, 64.4%,
91.7% and 81.2% of the trials were correctly classified with the EEG) and detection latencies within +/250ms with respect to the actual movements were obtained in most cases. The Fügl-Meyer Index and
the Stroke-Impact Scale were evaluated before and after the experimental month, and increases of
10.5+/-8.7 and 15.7+/-11.9 points were observed. Neurophysiological assessments of induced changes
with the intervention were carried out as well. The distribution of the cortical activation maps
remained stable across sessions in all cases. An increase in the amplitude of the readiness potential
peak (a possible biomarker of cortical excitability) was observed in two patients as a result of the
intervention. Finally, kinematic data of patients performing the reaching task were carried out before
and after the intervention. Results showed increased amplitudes in the shoulder extension during the
reaching phase in two of the patients and no changes in the other two. In summary, we functionally
tested an EEG-based associative facilitation therapy and observed a series of patient-specific changes
that could be associated with beneficial neuromodulation effects.
Poster Number: NM-138
Poster Name: Effect Deep Brain Stimulation of the Subthalamic Nucleus on Cortical and Subcortical
Neural Activity
Authorship List: Luke Johnson, Abirami Muralidharan, Jianyu Zhang, Matt Johnson, Kenneth Baker,
Jerrold Vitek
Institution: University of Minnesota
Abstract: Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an effective surgical
treatment for advanced Parkinson’s disease (PD). Although primary motor cortex (M1) is a key
component of the basal ganglia-thalamo-cortical network and likely plays a significant role in the
development of PD motor symptoms, few studies have characterized the firing patterns of M1 neurons
in the PD state or the effect of DBS on M1 activity. In this study we examined the effect of STN DBS on
spontaneous neuronal activity in M1 of a parkinsonian non-human primate. A 96-channel
microelectrode array was chronically implanted in the arm area of M1 to record single units (SU) and
local field potentials (LFPs), and an 8-contact macroelectrode was implanted in the STN to deliver
electrical stimulation and record LFPs. Spike activity of 79 single cells were simultaneously recorded
before, during, and after therapeutic STN DBS. Though DBS significantly altered the firing pattern of
96% of M1 cells, the median discharge rate of the total population was not changed significantly. The
prevalence of burst firing and oscillatory activity in the beta (13-30 Hz) range decreased modestly
during DBS. To assess cortical-subcortical synchronization, spike triggered averages (STAs) were
calculated by averaging the STN LFP signal around each recorded M1 spike. During DBS M1-STN
synchronization was significantly reduced, supporting the hypothesis that DBS works not just by
altering neuronal activity within nuclei of the basal ganglia, but also by changing the interactions
between nuclei within the basal ganglia-thalamo-cortical motor circuit.
Poster Number: NM-139
Poster Name: Improving Motor Recovery after Stroke by Combined rTMS and BCI Training
Authorship List: Nessa Johnson, Albert You, James Carey, Ann van de Winckel, Andrew Grande, Bin
He
Institution: University of Minnesota
Abstract: Treatment strategies to address motor impairment after stroke should optimally address
both contributors towards hemiparesis, namely by encouraging activity within the lesioned hemisphere
and down regulating inhibition from the healthy hemisphere. In this study, we sought to combine
repetitive Transcranial Magnetic Stimulation (rTMS) with motor imagery Brain Computer Interface
(BCI) training to enhance motor recovery after stroke. One stroke patient has completed the study to
date. Low frequency (1Hz) rTMS was applied to the motor hotspot of the non-stroke hemisphere at
90% motor threshold for 10 minutes, immediately followed by 8-10 runs of BCI training using a virtual
reality hand grasping task, with 20 trials per run. Nine combined rTMS/BCI sessions were completed
(three times per week for three weeks), followed by nine sessions of BCI training only. Clinical tests of
motor performance, paired-pulse TMS inter-hemispheric inhibition (IHI) tests, and functional MRI were
evaluated at four time points: baseline, post-rTMS/BCI, post-BCI, and follow-up. The subject was able
to achieve adequate control of the virtual reality BCI paradigm, with a peak percentage of 86% of
correct responses, and a daily average of nearly 70 % correct responses in the final BCI sessions.
Performance on a finger tracking test improved significantly over time, but no significant changes in
motor performance were seen in the Motricity or Box and Block Test. The results demonstrate the
feasibility of combining rTMS with BCI training in stroke patients, and lay a foundation for continued
work with additional subjects to evaluate the potential of this combined therapy.
Poster Number: NM-140
Poster Name: Mood Regulatory Actions of Nucleus Accumbens Deep Brain Stimulation
Authorship List: Rajas P. Kale, Abbas Z. Kouzani, Mark Frye, Ken Walder, Michael Berk, Susannah J.
Tye
Institution: Mayo Clinic, Deakin University (Australia), University of Minnesota
Abstract: DBS is shown to have antidepressant actions in treatment-resistant depression (TRD).
Nucleus accumbens (NAc) DBS have also been shown to induce mania in vulnerable individuals. This
study investigates the antidepressant effects of NAc DBS in an established animal model of TRD.
Wistar rats were divided into 4 groups: TRD-DBS (n=9), TRD-Sham (n=8), TRD (n=10), and Control
(n=10). Bilateral stimulating electrodes were implanted into the NAc of TRD-Sham and TRD-DBS
animals. Antidepressant-resistance and depression behaviors were induced through
adrenocorticotropic-hormone (ACTH-(1-24); 100µg/day; 2nd and 3rd weeks) administration and
concurrent social isolation (all 3 weeks) respectively. DBS was administered throughout the 2nd week
of ACTH treatment via a back mounted rodent DBS system. 24-hour locomotor activity counts were
obtained using infrared detectors and weekly sucrose preference tests were performed throughout
the 3 week protocol. Open field and FST were completed at the end of the 3 weeks. Brains were then
removed and stored at -80˚C. NAc tissue levels of brain-derived and glial-derived neurotrophic
factors (BDNF and GDNF, respectively) were quantified using western blot. Results demonstrate
significant increases in locomotor activity for TRD-DBS animals (DBS-Vs-Sham: p=0.0141). Lowered
immobility was observed during FST for TRD-DBS animals (DBS-Vs-Sham: p=0.0002). ACTH-induced
BDNF expression increased in the outer region substructure NAc-shell (p=0.0487) and decreased in
the inner region substructure NAc-core (p=0.0275) compared to controls. These data support
antidepressant actions of NAc DBS in TRD. Preliminary analysis of concurrent effects of daily
dopamine reuptake inhibitor GBR12909 (16mg/kg) administration coupled with NAc DBS demonstrates
dopamine-mediated augmentation of these behaviors.
Poster Number: NM-141
Poster Name: Principles of Within Electrode Current Steering (WECS)
Authorship List: Niranjan Khadka, Dennis Q. Troung, Marom Bikson
Institution: The City College of New York (CUNY)
Abstract: Within Electrode Current Steering (WECS), a novel method, applies to non-invasive
electrical stimulation with two or more electrodes to enhance reliability and tolerability during tDCS.
The underlying assumption of WECS is steering current within electrodes (to compensate for any nonideal conditions at the surface), but without altering current distribution in the brain target. This
technology leverages our technique for independently isolating electrode impedance and overpotential during multi-channel stimulation. Through an exemplary case example of a realistic
electrodes (metal-rivets embedded in an electrolyte (saline or gel)) and head geometry (FEM), we
demonstrated how current flow in brain is independent of current steering at the electrode. Three
current split cases (even, partially uneven, and fully uneven), keeping total current (1 mA) fixed within
the electrodes are tested. At the electrode-assembly interface with the skin, the current density
distribution varied only incrementally across conditions (e.g. less than would be expected) with even
minor changes in electrode assembly or skin properties. There was no difference in the predicted
electric filed at the brain target under all three cases. Thus, with such electrode assembly design,
current steering to any functional electrode would not significantly increase current density in the skin;
hence, not effecting tolerability.
Poster Number: NM-142
Poster Name: Vibro-tactile Stimulation as a Non-invasive Neuromodulation Method for Treating
Spasmodic Dysphonia
Authorship List: Sanaz Khosravani, Yu-Ting Tseng, I-Ling Yeh, Joshua Aman, Juergen Konczak
Institution: University of Minnesota
Abstract: Spasmodic dysphonia (SD) is a chronic neurological voice disorder characterized by
involuntary spasms of vocal cords. Currently, symptomatic treatment in form of botulinum toxin
injection is the most prominent therapy for SD. Vibro-tactile stimulation (VTS) has proven to
ameliorate motor manifestations in other types of focal dystonia. Here we first examined the unknown
effect of VTS on cortical activity in healthy volunteers and then applied VTS to SD patients to measure
the change in voice quality. Microvibrators were attached bilaterally to the skin above the thyroid.
Event related potentials in response to laryngeal vibration were recorded using a 64-channel EEG
system during vowel vocalization. The time-frequency analysis of the results revealed bilateral
suppression of cortical activity in somatosensory and motor areas in 2-20 Hz range i.e. α, β and μ bands
(suppression of μ is known to be associated with the intention to move). Second, three SD patients
participated in an assessment protocol comprising three sets of 15-minute stimulation: VTS-no
vocalization (5 min.) and vocalization with VTS (10 min.). Vocal quality was evaluated at the end of each
trial in both ‘on’ and ‘off’ vibratory conditions. Patient self-reports on voice quality and recorded audio
signals were analyzed a posteriori by a blinded rater using the Consensus Auditory-Perceptual
Evaluation of Voice (Cape-V). Results showed improvement of self-reported voice quality for all
participants and up to 35% improvement in average Cape-V scores. Both tests provide preliminary
evidence for the potential capability of using VTS as a non-invasive clinical treatment for spasmodic
dysphonia.
Poster Number: NM-143
Poster Name: Robotic Neuromuscular Facilitation for Regaining Neural Activation in Hemiparetic
Limbs
Authorship List: Jun Ueda, Lauren Lacey, Melih Turkseven, Minoru Shinohara, Ilya Kovalenko, Euisun
Kim, Fatiesa Sulejmani
Institution: Georgia Institute of Technology
Abstract: Poster presentation would introduce the author’s initial effort to understand and induce
functional recovery of hemiparetic limbs based on the concept of timing dependent neural plasticity.
Limb motor function is commonly impaired after neurologic injury such as stroke, with hemiparesis
being one of the major impairments. As an emerging unique intervention for hemiparesis, named
repetitive facilitation exercise, or RFE, a therapist manually applies brief mechanical stimuli to the
peripheral target muscles (e.g., tapping, stretching of tendon/muscle) immediately before a patient
intends to produce a movement with the muscle. The hypothesized mechanism of RFE is that stretch
reflex of mechanical stimuli overlaps with descending motor command and this takes advantage of
neural plasticity to strengthen neural pathway. While the practice of this rehabilitation procedure by a
skilled therapist often leads to dramatic rehabilitation outcome but not by unskilled therapists, most
likely due to the inaccuracy of the timing of peripheral stimulation in reference to the intention of
movement (i.e. motor command). MRI-compatible robotic rehabilitation device was designed to
improve the reliability and efficacy of the operation by satisfying the timing precision required by the
therapy. With the robotic device, this presentation would show the temporal dynamics of this
facilitation to understand the neuromuscular activation processes associated with RFE and further
improving RFE.
Poster Number: NM-144
Poster Name: Robust Prediction of Clinical Deep Brain Stimulation Targets
Authorship List: Jinyoung Kim, Yuval Duchin, Guillermo Sapiro, Jerrold Vitek, Noam Harel
Institution: Duke University, University of Minnesota, SIS
Abstract: This work presents a robust prediction framework that enables direct targeting of
anatomical structures for Deep Brain Stimulation (DBS) surgery based on standard clinical MRI (1.5
Tesla). With recent advances in ultrahigh-field MR (7T), the superior contrast and high resolution
imaging allow direct visualization and localization of the DBS targets on the individual patient.
However, such ultrahigh-fields are not clinically available, and therefore DBS targeting needs to be
performed on the standard clinical low contrast data. Unfortunately, clear visualization of DBS targets
is not feasible with such standard clinical MRI protocols. To address the DBS targeting problem on the
clinical 1.5T MRI, we first learn the shape relationships between DBS targets and their potential
predictors from ultrahigh-field (7 Tesla) MR training sets via a bagging regression model, thereby
reducing the variability of learned dependencies. Then, given manually or automatically detected
predictors on the clinical patient data, the target structure is predicted using the learned high quality
information. Furthermore, we derive a robust way to properly weight different training subsets when
using an ensemble of predictions, based on the contribution to the prediction estimated via a machine
learning approach. The subthalamic nucleus (STN) which is critical for Parkinson's disease (PD) and
potentially also for obsessive compulsive disorder is used to exemplify within our proposed
framework. Experimental results from PD patients validate that the proposed approach provides
reliable prediction of the STN from the clinical 1.5T MR data.
Poster Number: NM-145
Poster Name: TMS of Laryngeal Muscles to Investigate Cortical Excitability in Spasmodic Dysphonia
Authorship List: T. J. Kimberley, M. Chen, R. L. Schmidt, K. O. Lim, J. Camchong
Institution: University of Minnesota
Abstract: Adductor spasmodic dysphonia (AdSD) is a focal dystonia affecting the laryngeal muscles
during speech. The pathophysiology of AdSD is unclear. Transcranial magnetic stimulation (TMS)
studies on the excitability of the laryngeal M1 connections to affected muscles [thyroarytenoid, (TA)],
has not been investigated due to difficulties secondary to the intrinsic nature of the TA muscle and in
accessing/confirming the laryngeal motor cortex (LMC). Furthermore, how the TMS ‘hotspot’ relates
to the fMRI voluntary motor task peak activation area is also unknown. This study determines the
cortical excitability and the TMS hotspot location using the TA muscle EMG responses to TMS over
the LMC. It maps the location of the identified TMS hotspot and the peak BOLD task-fMRI activation
cluster onto the rest fMRI data. The strength of resting functional connectivity between these two
identified regions is compared between AdSD and healthy controls. Preliminary data of 9 subjects (5
AdSD, 4 healthy, age matched) will be presented. The results show that: the fine wire electrode EMG
responses can be used to assess LMC excitability with the CSP. Cortical excitability results may
suggest longer CSPs in the AdSD group than the HC group, which may indicate a more inhibited LMC
in AdSD subjects. Connectivity data may suggest lower connectivity in the LMC in AdSD group than
that of the HC group. This may indicate a mismatch between the voluntary functional network (taskfMRI BOLD signal) and the passive structural network (TMS) in the laryngeal area of the AdSD.
Poster Number: NM-146
Poster Name: Optogenetic Neuromodulation for Temporal Lobe Epilepsy
Authorship List: Esther Krook-Magnuson
Institution: University of Minnesota
Abstract: On-demand optogenetic techniques provide temporal, spatial, cell-type, and direction of
modulation specificity, allowing “where you need it, when you need it” intervention. Using these
techniques in a mouse model of temporal lobe epilepsy, we demonstrated: (i) broad inhibition of
excitatory cells in the hippocampus can truncate seizures1; (ii) selective inhibition of granule cells in the
dentate gyrus ipsilateral to sclerotic tissue truncates seizures2; (iii) excitation of parvalbumin positive
hippocampal interneurons ipsilateral or bilateral to sclerotic tissue truncates seizures1; (iv) inhibition of
electrographic seizures reduces the frequency of behavioral seizures1; and more recently, (v) that the
cerebellum can be an effective target for intervention for temporal lobe seizures3. Interestingly,
inhibition of temporal lobe seizures through cerebellar directed intervention was robust, occurring
with excitation or inhibition of Purkinje cells, with manipulation of the midline or lateral cerebellum,
and with short or long light pulses. With excitation of the midline cerebellum uniquely there was an
additional effect on the time to next seizure. We further found that temporal lobe seizures modulated
the activity of cerebellar neurons3. Future studies in my lab will investigate this bi-directional
functional connectivity and the mechanisms underlying cerebellar control of temporal lobe seizures.
Funding: NINDS NS087110
References:
1 Krook-Magnuson, E., Armstrong, C., Oijala, M. & Soltesz, I. Nature communications (2013).
2 Krook-Magnuson, E., Armstrong, C., Bui, A., Lew, S., Oijala, M. & Soltesz, I. J Phsyiol (in press).
3 Krook-Magnuson, E., Szabo, G. G., Armstrong, C., Oijala, M. & Soltesz, I. eNeruo (2014).
Poster Number: NM-147
Poster Name: tDCS to Enhance Therapeutic Learning in Addicted Individuals: A Novel Translational
Model
Authorship List: Matt G. Kushner, Kelly Berg, Justin Anker, Paul Thurus, Marc Mooney, Amanda Unruh
Institution: University of Minnesota
Abstract: Transcranial direct current stimulation (tDCS) is a non-invasive neuro-modulation technology
that is considered to have the potential for translation as a psychiatric treatment. The model reflected
in this nascent literature envisions tDCS as a primary intervention to remediate or compensate for a
presumptive neurological defect underlying a particular mental illness or symptom manifestation;
however, at this point, the biological defect, the mechanism of change and the ultimate effectiveness
of such an intervention all remain speculative. Here we propose a novel translational model in which
the known memory/ attention enhancing effects of tDCS would be leveraged to amplify the retention
and benefit of therapeutic learning associated with established cognitive and behavioral therapy (CBT)
procedures. We have been awarded a small “start-up” grant to pilot this approach (Berg, PI) within our
ongoing NIH R01 (Kushner, PI) in which recently detoxified alcohol dependent individuals are given a
CBT that teaches specific coping skills aimed at maintaining sobriety. As a proof-of-concept, 10 to 20
cases form the R01 will undergo 20 minutes of tDCS anodal stimulation targeting the left DLPFC (area
typically stimulated in published tDCS memory enhancement studies) or sham tDCS immediately
following each of the six CBT sessions. Primary outcome measures will be free recall and multiple
choice tests of memory for therapy content given 24 hours following the stimulation. It is predicted
that those receiving tDCS will have better recall of CBT material than do controls. A secondary
analysis will examine whether the active tDCS group experiences better clinical outcomes 1 and 4
months following treatment and specialized statistical tests will be used to help disambiguate direct
versus mediated tDCS effects. Although results of this study will not be available by the time of the
meeting, this poster is meant to provide a detailed example of this translational neuromodulation-CBT
hybrid concept to interested researchers and clinicians. We argue that this approach represents a
significant conceptual innovation that expands the current conceptual horizon envisioned for
translating tDCS to psychiatric applications. Beyond alcoholism, this novel tDCS translational
application could potentially enhance treatment outcomes in a variety of otherwise sub-optimally
treated psychiatric conditions; especially those with a significant behavioral component (e.g.,
addictions, eating disorders and “habit” disorders such as pathological gambling).
Poster Number: NM-148
Poster Name: Determination of Stimulation Focality in Heterogeneous Head Models During
Transcranial Magnetic Stimulation (TMS)
Authorship List: Erik Lee, Ravi Hadimani, David Jiles
Institution: Iowa State University
Abstract: Transcranial Magnetic Stimulation (TMS) is an increasingly popular tool used by both the
scientific and medical community to understand and treat the brain. TMS has the potential to help
people with a wide range of diseases such as Parkinson's Disease, Obsessive Compulsive Disorder,
and Post Traumatic Stress Disorder, while currently being used to treat people with chronic, drugresistant depression. Through computer simulations, we are able to see the electric field that TMS
induces in anatomical human models, but there is no measure to quantify this electric field in a way
that relates to a specific patient undergoing TMS therapy. We propose a way to quantify the focality of
the induced electric field in a heterogeneous head model during TMS by relating the surface area of
the brain being stimulated to the total volume of the brain being stimulated. This figure would be
obtained by conducting finite element analysis (FEA) simulations of TMS therapy on a patient specific
head model. Using this figure to assist in TMS therapy will allow clinicians and researchers to more
accurately stimulate the desired region of a patient's brain and be more equipped to do comparative
studies on the effects of TMS across different patients.
Poster Number: NM-149
Poster Name: The Use of Vagus Nerve Stimulation to Treat Cardiovascular and Metabolic Diseases
Authorship List: Steven W. Lee, Elizabeth M. Annoni, Xueyi Xie, Kanchan Kulkarni, Bruce H.
KenKnight, John W. Osborn, Elena G. Tolkacheva
Institution: University of Minnesota
Abstract: Vagal nerve stimulation (VNS) is a neuromodulation therapy that involves delivering
electrical pulses to the vagus nerve via an implanted system. VNS is an FDA-approved treatment for
intractable epilepsy and treatment-resistant depression. However, the potential use of VNS as a
therapy to treat cardiovascular diseases remains unclear. The aim of this therapy is to modulate the
parasympathetic activity in order to reestablish the cardiac autonomic balance in response to the
increased sympathetic tone seen in cardiovascular diseases. In our laboratory, we investigated the
therapeutic potential of VNS on two different cardiac disease models: 1) hypertension (HTN) and 2)
myocardial infarction (MI).
To study the effects of VNS, both in vivo and ex vivo studies were performed. Continuous
cardiovascular data acquisition was performed in vivo to obtain real-time ECG and blood pressure
data. Ex vivo optical mapping studies were performed to study the electrophysiological remodeling
that occurs in both diseased and VNS-treated hearts.
In our HTN model, after 4 weeks of VNS treatment, both in vivo and ex vivo studies showed a
significant decrease in the number of arrhythmia episodes in HTN-VNS rats compared to the HTNSham rats. The ex vivo optical mapping results clearly showed electrical remodeling of the heart due to
VNS. In the HTN-VNS rats, the conduction velocity increased, the heterogeneity in the ventricles
decreased, and the action potential duration decreased in the RV, which suggests that the heart is less
arrhythmogenic.
Hence, our results suggest that VNS can be used as a potential therapy for cardiac diseases.
Poster Number: NM-150
Poster Name: Local Electrical Properties Tomography with Global Regularization by Gradient
Authorship List: Jiaen Liu, Xiaotong Zhang, Yicun Wang, Pierre-Francois Van de Moortele, Bin He
Institution: University of Minnesota
Abstract: The in vivo electrical conductivity and permittivity of brain tissues can provide important
information regarding the fundamental status of brain function and health condition, and guiding
neuromodulation therapy of brain disorders. Recently, the electrical properties tomography (EPT)
method has been introduced to image these electrical properties in vivo based on the measured
radiofrequency (RF) magnetic field in MRI. However, the traditional EPT methods have been suffered
from measurement noise contamination and overly simplified central equations that fail in regions of
inhomogeneous electrical properties, e.g. near boundaries of tissues. Although a gradient-based EPT
method has been developed previously to address these limitations, it requires measurements of both
transmit and receive RF fields and relies on an impractical assumption requiring symmetrical samples.
In this study, a new algorithm was proposed to improve the performance near tissue boundaries by
constraining the global electrical-properties maps regularized by the calculated incomplete gradient
information solely using transmit RF field. Results from experiments of a physical phantom and brains
of healthy human subjects suggest that the proposed algorithm has significantly improved imaging
accuracy and spatial resolution while significantly reducing scan time and broadening its application on
subjects with an asymmetric head.
Poster Number: NM-152
Poster Name: Image-guided Stereotactic Frame for Intraspinal Microstimulation Delivery
Authorship List: Aldo A. Mendez, Peter J. Grahn, James Trevathan, Grant W. Mallory, Steve J. Goerss,
Ju Ho Jeong, Darlene A. Lobel, Allan J. Bieber, Bruce A. Kall, Bryan L. Striemer, Joel P. Felmlee, Kevin
E. Bennet, Luis Lujan, Kendall H. Lee
Institution: Mayo Clinic
Abstract: Animal studies have shown that Intraspinal microstimulation (ISMS), an emerging technique
for evoking limb movements following spinal cord injury, is capable of controlling limb movements
while improving fatigue resistance. A significant obstacle to clinical translation of ISMS is the difficulty
of targeting spinal cord regions responsible for selective motor function. Presently, electrode targeting
relies on external anatomical landmarks, and is thus susceptible to targeting errors due to anatomical
differences between subjects.
We have designed a magnetic resonance imaging (MRI)-guided stereotactic targeting and electrode
delivery system. This system offers the potential for reducing the spatial inaccuracy of existing
methods, thereby improving targeting and selective stimulation of neuronal populations responsible
for specific motor function.
We verified the accuracy of the image-guided stereotactic system using anthropomorphic phantoms
and pig cadavers. Initial application of the stereotactic system is intended for ISMS. However, the
targeting and delivery system described herein is applicable to other therapeutic applications such as
fine needle biopsies and intraspinal delivery of drugs and stem cells.
Poster Number: NM-153
Poster Name: Different Learning Processes of Multichannel versus Small Channel Configuration for
Online Brain-computer Interface
Authorship List: Jianjun Meng, Shuying Zhang, Angeliki Beyko, Taylour Hanson, Bin He
Institution: University of Minnesota
Abstract: Motor imagery based brain computer interface (BCI) using electroencepholography (EEG)
has shown promising results in control of virtual objects, which can be used to modulate brain states.
However, only a few subjects have demonstrated good control with C3 and C4 electrodes positioned
above the motor cortex. Typical BCI studies incorporate a small channel configuration, known as a
large Laplacian, which uses the same two electrodes/channels together with surrounding channels. A
more recent approach is the multichannel configuration, which can utilize more EEG electrodes than
the small channel configuration. With additional electrodes and different online decoding algorithms,
better performance and new learning processes may be expected. One challenge is that additional
channels lead to higher inclusion of both task-related and -unrelated information, and thus may
obfuscate the task-related information and affect the efficiency and stability of learning processes.
This study aimed to compare the learning processes between the two configurations (multichannel vs.
small channel) by evaluating the average online performance across several sessions. The experimental
results revealed that the multichannel configuration exhibited significantly better performance initially
within naïve subjects, though did not lead to a congruent improvement in long-term learning. In
contrast, the small channel configuration displayed a slow and steady improvement on average across
all sessions. The results indicate that learning may initially involve a larger brain area; whereas, as
learning progresses, the motor cortex becomes more crucial. This might explain why the multichannel
configuration can decode motor imagination better initially, while small channel improved after a longterm learning period.
Poster Number: NM-154
Poster Name: A Noninvasive Neuromodulation Treatment for Tinnitus Using a Paired Paradigm
Authorship List: Yezihalem Mesfin, Jennifer K. Wittman, Cory D. Gloeckner, Craig D. Markovitz,
Hubert H. Lim
Institution: University of Minnesota
Abstract:
Objectives: We propose a noninvasive approach (mSync) for treating tinnitus. Our previous animal
studies showed that combining broadband noise and somatosensory electrical stimulation elicits
differential plasticity in auditory regions linked to tinnitus. Greater neural suppression versus
facilitation was also achieved using specific body locations (e.g., right ear) and precise inter-stimulus
intervals (e.g., body 5 ms before noise), which may suppress the hyperactivity linked to tinnitus. We
now plan to investigate the effects of different mSync parameters on tinnitus in human subjects. Since
~5% of the population suffers from tinnitus, mSync could have a major clinical impact in society.
Methods: In 10 sessions, we will investigate how surface electrical stimulation of different body sites
(left ear, right ear, tongue, upper back/neck, left hand or right hand) with varying delays (-25, -15, -5, +5,
15 or 25 ms) between a single electrical pulse on the body and a broadband acoustic stimulus can
interact with and decrease tinnitus. In 5 sessions, we will add single pulse transcranial magnetic
stimulation of the frontal cortex with varying delays (0, +10, +20 or +30 ms) relative to body-acoustic
stimulation to attempt to further reinforce the mSync effects.
Expected Plans: We recently obtained IRB approval. We will recruit 20 tinnitus subjects and begin
testing in April 2015. We will explore different mSync parameters and characterize the corresponding
changes in the tinnitus percept using various tests and questionnaires. Success with this pilot study will
motivate a larger clinical study to improve mSync.
Funding: University of Minnesota’s Institute for Engineering in Medicine and MnDRIVE
Poster Number: NM-155
Poster Name: Spinal Cord Stimulation in an Ovine Model of Neuropathic Pain Measured through Von
Frey Filaments, Gait Analysis, and Dorsal Horn Recordings
Authorship List: John W Miller, Chandan G Reddy, Saul Wilson, Brian D Dalm, Sina Safayi, Sara K.
Shivapour, Kingsley Abode- Iyamah, Stephanus Viljoen, Douglas C Fredericks, Katherine N GibsonCorley, Nicole M Grosland, Nivedita U Jerath, Kirsten Stoner, Richard Reale, Hiroyuki Oya, Nicholas D
Jeffery, Timothy J Brennan, George T Gillies, Matthew A Howard III
Institution: University of Iowa Hospitals and Clinics, University of Iowa, University of Virginia
Abstract: We have developed an ovine model of neuropathic pain intended for use in investigating the
efficacy and possible mechanisms of action of spinal cord stimulation. Chronic constriction, i.e., 25%
reduction in diameter, of the peroneal nerve produces quantifiable behavioral and neurophysiological
responses, without the deficits associated with complete ligation, which seem to be modulated by
spinal cord stimulation. In particular, we observed that epidural stimulations from 0.1 to 0.5 V result in
an apparent increase in tolerance to von Frey filaments applied to the affected limb, recovery to nearnormal hoof elevation during treadmill ambulation, and preliminary evidence from gait analysis
suggesting acclimation to the stimulation signal within < 150 steps on the treadmill. In parallel with this,
we have developed neurophysiological methods for obtaining single unit recordings from dorsal horn
neurons identified as being those most responsive to the stimulation signals, and for making
electromyographic measurements of the hind limb muscles affected by the nerve constriction. We
discuss the next steps in the ongoing development and application of this large animal model, and the
implications of our pilot study findings in terms of their impact on the development of a novel
intradural form of spinal cord stimulation.
Poster Number: NM-156
Poster Name: Training Produces Enhanced Neural Sensitivity for Speech Perception in Adult
Cochlear Implant Users
Authorship List: Sharon Miller, Yang Zhang, Peggy Nelson
Institution: University of Minnesota
Abstract: Postlingually deafened cochlear implant (CI) users have to learn to remap degraded
electrical stimulations to previously learned speech sounds and patterns. Recent studies have shown
that targeted auditory training can improve speech and music perception, but the neural mechanisms
that support these improvements remain unclear. A pilot study in our lab confirmed that targeted
phoneme training utilizing multi-talker variability significantly improved identification of naturally
produced /ba/, /da/, /wa/, and /ya/ stimuli in adult CI users and the improvement generalized to
novel speech stimuli. The purpose of the present study was to better understand the neural basis of
the improved phoneme identification. We examined whether our training protocol could induce more
categorical-like perception of the trained phonemes in CI users. We measured phoneme identification
and discrimination and mismatch negativity (MMN) responses to within- and across- category speech
and nonspeech contrasts before and after training from 9 postlingually deafened adult CI users. The
results showed that training significantly increased MMN amplitudes to the across category phoneme
contrast and no significant effects on the nonspeech control stimuli. MMN amplitudes to the withincategory phoneme contrast did not significantly change after training for either speech or nonspeech
contrasts. Changes in across-category MMN amplitude for speech stimuli were significantly correlated
to changes in the phoneme identification boundary. Together, the results demonstrate the feasibility
of intensive laboratory training to improve cochlear implant users' speech perception and the utility of
the MMN as a clinical tool to assess underlying changes in neural sensitivity to speech sounds.
Poster Number: NM-158
Poster Name: V4 LFP Signals Predict and Affect Behavioral Reliability in Non-human Primates During
Shape Detection Task Training
Authorship List: Elisabeth Moore, Katherine Weiner, Geoffrey Ghose
Institution: University of Minnesota
Abstract: Learning a new ability requires a change in the representations present in cerebral cortex,
and visual perceptual training is a useful paradigm for studying learning, as neurons within the visual
system can be quantitatively characterized. We hypothesize that area V4 neurons reflect both the
visual stimulus and behavioral choice in a shape detection task, and are responsible for associated
performance improvements.
We recorded in V4 from a 96-electrode array, while two non-human primates learned a shape
detection task. Behavioral reliability (performance) improved during training for both animals,
indicating learning. LFP information was computed in varying windows of time, at varying delays since
sensory (stimulus appearance) or behavioral choice event. We combined sensory and choice
information to assess whether LFPs predict behavior and found that small numbers of electrodes can
predict behavioral timing and reliability. Daily variations in electrode predictions were also correlated
with behavior, with opposing correlation signs for the two animals.
These results suggest that V4 plays a critical role in shape detection and that performance
improvements can be explained by changes within local populations of V4 neurons. To conclusively
establish this, we paired microstimulation in V4 with visual training to improve learning outcomes.
Preliminary data suggests that lasting performance changes are possible, but that the sign of the
change depends on small variations in the spatial and temporal distribution of microstimulation
delivered. Thus, with respect to neural interface design, maximal efficacy is likely to require a precise
match of stimulation parameters to the signals present during normal behavior.
Poster Number: NM-159
Poster Name: Developing Adaptive Stimulation Algorithms to Modulate Hippocampal and Seizure
Dynamics
Authorship List: Vivek Nagaraj, Theoden Netoff
Institution: University of Minnesota
Abstract: For many patients with Epilepsy, antiepileptic drugs do not fully control seizures. Deep Brain
Stimulation (DBS) devices can provide some benefit to patients; however, efficacy can be improved
through understanding how DBS modulates neural networks with low seizure thresholds. This research
proposes a novel closed-loop stimulation algorithm that adaptively modulates stimulation parameters
with respect to salient physiological signals. The goal of this research is to suppress seizures while
minimizing total stimulation energy.
Poster Number: NM-160
Poster Name: Minimally-invasive Optogenetics in the Non-human Primate
Authorship List: Jonathan J. Nassi, Michael C. Avery, Ali H. Cetin, Anna W. Roe, John H. Reynolds
Institution: The Salk Institute for Biological Studies, Vanderbilt University
Abstract: Targeted modulation of neural activity in the central nervous system can alleviate symptoms
of neurological disease in humans. However, current therapies, such as electrical stimulation, are highly
invasive and of limited long-term efficacy due to their inability to target specific cell-types.
Optogenetics is a powerful new technique that allows for temporally-precise, cell-type specific control
of neural activity, with the potential to significantly improve clinical outcomes. Optogenetics has
already been successfully applied in the rodent to probe the neural mechanisms underlying behavior
and disease, but its application in the non-human primate (NHP), a critical step toward clinical
application in humans, has been slow to develop. One key challenge has been the delivery of viruses
and light to the brain, which, in primates, must pass through a thick and opaque dura mater that
overlies the cortex. Here, we report optogenetic modulation of neural activity in the alert and
behaving macaque after replacement of the native dura with a transparent artificial dura. This
approach enables transdural illumination which obviates the damage that would otherwise occur as a
result of lowering optical fibers into the brain. We have used this newly-developed technique to
investigate how optogenetic neuromodulation interacts with natural patterns of cortical activity
following visual stimulation and changes in cognitive state. This experimental paradigm promises to
greatly assist in the dissection of cortical circuits underlying perception, cognition and behavior in the
NHP, and should greatly advance ongoing efforts to translate this exciting new technology into a viable
therapy for human neurological disease.
Poster Number: NM-161
Poster Name: Transcranial Current Stimulation for the Treatment of Medication Refractory Auditory
Hallucinations
Authorship List: Brent G. Nelson, Suzanne G. Jasberg, Carolyn Gentz, Molly Gierke, Ben Otopalik,
Casey Gilmore, Jazmin Camchong, Kathryn Cullen, S. Charles Schulz, Kelvin O. Lim
Institution: University of Minnesota
Abstract: Auditory hallucinations (AH) in schizophrenia are medication refractory in 30% of patients.
Transcranial current stimulation (tCS) is promising, consisting of transcranial direct current (tDCS) and
transcranial random noise (tRNS) stimulation. Both involve applying weak electrical current (<2mA) to
scalp enhancing cortical excitability. Brunelin utilized tDCS to reduce medication refractory AH and
Vanneste applied tRNS to reduce loudness and distress associated with tinnitus. Our study compares
the efficacy of tRNS to tDCS for medication refractory AH.
This double-blind, sham-controlled study includes three arms (tRNS, tDCS, and sham; DLPFC and
temporal parietal junction). Two treatment sessions/day for 5 days. Assessments include Positive and
Negative Symptoms Scale (PANSS) and Auditory Hallucinations Rating Scale (AHRS), completed
before the 1st and after the last treatment, and at 1, 3, 6, 9, and 12 month follow-up.
During a piloting phase, 3 subjects were treated with tDCS and analyzed through 1 month. Subject 1:
AHRS increases in frequency and salience, decreases in length and loudness (total score increase of 4)
and PANSS total decrease of 2. Subject 2: AHRS decreases in all domains (total score decrease of 18)
and PANSS total decrease of 26. Subject 3: AHRS increase in distress but decreases in salience,
frequency, and loudness (total score decrease of 2) and PANSS total decrease of 2.
Our data suggests subjects respond dirrerently to tDCS, possibly due to baseline nature of
hallucinations. Loudness decreased in all subjects, thus specific sub symptoms may be identified.
Additional subjects will be treated and analyzed at a group level.
Poster Number: NM-162
Poster Name: Electrochemical Method for Real-time Detection of NADH and NAD
Authorship List: Evan Nicolai, Aldo Mendez, Su-youne Chang
Institution: Mayo Clinic
Abstract: Nicotinamide adenine dinucleotide (NADH/NAD) plays an important role in metabolism and
it was recently discovered that NAD might have a neuromodulatory function. To define its
neuromodulatory and physiological function, it is important to monitor these molecules in real-time.
Fast-scan cyclic voltammetry (FSCV) is an electrochemical method that has been used for real-time
monitoring of electroactive molecules including dopamine, adenosine, serotonin, histamine, and
hydrogen peroxide. Here, we have defined a way to detect NADH/NAD using FSCV at the tip of a
carbon fiber microelectrode. The optimal parameter for NADH and NAD is a triangle waveform
ramping from -0.4 to 1.45 V and back to -0.4 V at a scan rate of 600 V/s at a repetition rate of 10 Hz.
Three oxidations were obtained at +0.8, +1.2, and +1.3 V, and one reduction at 0 V for NADH. The
oxidation at 0.8 V is greatly diminished in recordings of NAD; however, other redox potentials are
similar to NADH. The determined optimal parameter was also applied to sense adenosine, dopamine,
and nicotinamide. For adenosine, two oxidation peaks were found at +1.3 and +1.45 V with no reduction.
For dopamine, one oxidation peak was found at +0.8 V and one reduction peak at -0.3 V. The FSCV
protocol was insensitive to nicotinamide. Therefore, our method is able to discriminate NADH/NAD
from molecules that have similar redox potentials. The ability to measure NADH and NAD using a bare
carbon fiber microelectrode in real-time will provide new biological insight into these two
tremendously important molecules.
Poster Number: NM-163
Poster Name: Investigating a New Tinnitus Treatment using the Auditory Midbrain Implant
Authorship List: Sarah J. Offutt, Robert J. Hughes, Hubert H. Lim
Institution: University of Minnesota
Abstract: Tinnitus is a phantom auditory percept that is disturbing and even debilitating for millions of
individuals in the US alone. Tinnitus has commonly been associated with properties such as
hyperactivity and increased neural synchrony across the auditory system, and thus one method of
treatment is to suppress this tinnitus-related activity to possibly reduce the tinnitus percept. We
propose a new treatment option that utilizes this method by delivering deep brain stimulation to the
inferior colliculus, which can potentially modulate the tinnitus-related activity across the auditory
system. In this study, we investigated the potential efficacy of this treatment by examining the
modulatory effects of stimulation of the dorsal cortex of the inferior colliculus on the central nucleus
of the inferior colliculus. Neural modulation and synchrony were assessed in response to electricalonly and paired acoustic-electrical paradigms with varying time delays, and compared to modulation
achieved by control paradigms. The effects were assessed immediately and 30 minutes following the
stimulation paradigms. Our results reveal that overall we get more suppressive that facilitatory
modulation and that significant, long lasting suppression can be observed for the paired paradigm with
an 18 ms delay and the electrical-only paradigm. No changes were observed for synchrony. Based on
our findings, these two paradigms could potentially provide lasting suppression of the hyperactivity
associated with tinnitus and this treatment will be tested in an upcoming clinical trial using the auditory
midbrain implant.
Poster Number: NM-164
Poster Name: Does an Intraneural Interface Short-term Implant for Robotic Hand Control Modulate
Sensorimotor Cortical Integration? An EEG-TMS Co-registration Study on a Human Amputee
Authorship List: S. Petrichella, A. Guerra, F. Ferreri, D. Ponzo, L. Vollero, G. Di Pino, A. Benvenuto, M.
Tombini, L. Rossini, L. Denaro, S. Micera, G. Iannello, E. Guglielmelli, V. Denaro, P.M. Rossini
Institution: University Campus Bio-Medico (Rome, Italy)
Abstract: Due to a discrete number of survived fibers and to the redundancy of connections and
despite plastic modifications at all nervous system levels, amputations makes the peripheral nerve
connections functionally silent or involved in some “aberrant” plastic reorganization. Aim of this study
is to investigate, using an EEG-TMS co-registration study, how direct bidirectional connection between
brain and hand prosthesis modifies the sensorimotor cortical integration.
A 26 years-old, left-hand amputated male have implanted four intrafascicular electrodes (tf-LIFEs-4)
two for each nerve, in the median and ulnar nerves of the stump for 4 weeks. Before tf-LIFE-4s implant
(T0) and after the training period, during which was employed a stand-alone version of the CyberHand
prototype, which approximates dimensions and grasping capabilities of the human hand with five
fingers actuated by six motors, once electrodes have been removed (T1), the subject underwents
neuronavigated EEG–TMS experiment.
In the right hemisphere, contralateral to the stump, it was possible to observe at T0 a slight increase
and anterior displacement of the positive pole of the dipole classically centered on the stimulated M1
around 46 ms and at T1 a general reduction of cortical excitability that remarkably involved the contralateral prefrontal cortices. This reduction was evident in most of the TMS-evoked potentials elicited by
the stimulation of the right hemisphere (N46 and P60 ipsilateral and P30 and N100 contralateral).
The results of this study confirm the hypothesis that bidirectional neural interface could redirect
cortical areas that the amputation have deprived of their original function, toward restorative
neuroplasticity.
Poster Number: NM-165
Poster Name: SSVEP Signatures in Visual Rivalry
Authorship List: Vadim Petruk, Steve Engel, Bin He, Sheng He
Institution: University of Minnesota
Abstract: In visual competition, two distinct perceptual representations compete for access to
awareness, however, it is unknown where this competition takes place. Previous studies suggest a lowlevel feature competition within monocular neural representations, or alternatively a high-level
stimulus competition within binocular neural representations. We used the electroencephalogram
(EEG) to investigate steady-state visually evoked potentials during visual rivalry either engaging high
level or low level competition. The high level competition condition included monocular and stimulus
rivalry whereas low level competition condition was binocular rivalry. We tagged each stimulus with a
frequency tag such that we could monitor the neural processing related to each stimulus specifically.
Source localization on the difference in amplitudes of the SSVEP frequency tags indicated V1
activation in all types of rivalry competition. Thus, visual rivalry may involve a hybrid of both high level
and low level neural representation competition. A multilayer model of interocular competition in layer
4 and cross orientation suppression in layer 2/3 is presented as a potential neural mechanism
mediating competitive interactions involved in visual rivalry.
Poster Number: NM-166
Poster Name: Cutaneomuscular Reflex Neuromodulation During Rest and Short-term
Neuroneuromodulation Induced by Leg-cycling after Spinal Cord Injury
Authorship List: Piazza, Taylor, Gómez Soriano, Bravo-Esteban, Torricelli, Avila-Martin, Pons
Institution: Spanish National Research Council (CSIC) (Madrid, Spain), National Spinal Cord Injury
Hospital (Toledo, Spain), Castilla La Mancha University (Toledo, Spain), Zaragoza University (Zaragoza,
Spain)
Abstract:
Objective: To characterise cutaneomuscular H-reflex neuromodulation during rest and its
neuroplasticity following a 10-minute leg cycling session during subacute spinal cord injury.
Study Design: A transverse descriptive study in non-injured subjects (n=10), and incomplete (iSCI, n=9)
individuals with subacute spinal cord injury.
Setting: Hospital Nacional de Parapléjicos, Toledo, SESCAM, Spain.
Methods: Soleus H-reflex neuromodulation with an ipsilateral cutaneomuscular plantar electrical
stimuli applied at 25, 50, 75 and 100ms inter-stimulus intervals (ISI´s) assessed during rest, and also
neuroplasticity induced after a 10-minute leg-cycling session (42rpm) in the non-injured and motor iSCI
groups.
Results: H-reflex 25ms ISI inhibition (-28±4%) observed in the non-injured group was significantly
reduced in the iSCI (-12±4%) group, while late 100ms ISI excitability (20±4% in the non-injured group)
was absent. After ten-minute cycling short ISI H-reflex excitability increased in the non-injured group
(20±8% for 25ms and 17±8% for 50ms ISI), while reflex activity was inhibited at 75ms ISI in the iSCI
group (-14±9%). However, exercise-induced H-reflex neuroplasticity at 75ms ISI in the iSCI group
correlated both positively (r=0.85) and negatively (r=-0.85) with the Triceps Surae muscle and distal
lower-limb hypertonia scores, respectively.
Conclusions: Leg-cycling promotes short-term spinal reflex neuroplasticity in the iSCI group which is
related to residual muscle function. Analysis of cutaneomuscular-evoked H-reflex neuromodulation
induced by leg-cycling highlights individual rehabilitation response profiles in the context of adaptive
and maladaptive spinal motor neuroplasticity.
Sponsorship: Project funded by the HYPER (CSD2009-00067) and Fundación Mutua Madrileña
(2013).
Keywords: Plantar cutaneomuscular stimulation; Soleus H-reflex; Reflex neuroplasticity; Passive and
assisted leg-cycling; Manual muscle score; SCI spasticity syndrome
Poster Number: NM-167
Poster Name: Functional Connectivity in the Visual Cortex using Cortical Depth Dependent fMRI
Authorship List: Cheng Qiu, Daniel Kersten, Cheryl A. Olman
Institution: University of Minnesota
Abstract: The cerebral cortex has a laminar structure with six major horizontal layers spanning a depth
of 2-4 mm. Projections from lower-level to higher-level cortical areas arise mainly from superficial
layers and terminate in middle layers; projections from higher-level back to lower-level areas arise
from both superficial and deep layers and terminate outside the middle layers. Understanding brain
functions and connectivity through cortical depth has the potential to identify effective targets for
neuromodulation.
In this study, we used high field fMRI with submillimeter resolution to explore cortical depthdependent responses and their correlations in the human visual cortex. Subjects performed a task
engaging both the ventral and dorsal visual areas: detecting the target object from a series of 3D
structure-from-motion stimuli whose coherence level was systematically modulated in blocks.
Functional imaging data (0.8 mm isotropic voxels, 4s TR) were collected using a 7T Siemens scanner
with Spin Echo EPI, and they were sorted into multiple equally populated depth bins. We found that
the stimulus coherence level significantly modulated BOLD responses in intermediate and higher-level
visual areas. In the comparison of correlations between average time series from those depth bins,
early visual areas V1 and V2, shape sensitive area LOC, and motion sensitive areas V3AB and hMT+
showed strong intra-area correlations (between depth bins within each area); inter-area correlations
(between depth bins from different visual areas) among early visual areas or between hV4 and LOC
are positive, whereas areas V1 and LOC are negatively correlated.
Poster Number: NM-168
Poster Name: Determining the Location of the Primary Motor Cortex in Children: A Comparison of
Electroencephalogram (EEG) and Transcranial Magnetic Stimulation (TMS) Techniques
Authorship List: Tonya Rich, Edgar Peña, Tim Feyma, Gregg Meekins, Nikolas Sell, Bernadette Gillick
Institution: University of Minnesota
Abstract: Transcranial Direct Current Stimulation (tDCS) is an emerging intervention to potentiate
motor recovery in children with hemiparesis. Traditionally, electrode placement targeting the primary
motor cortex for hand representation has relied on the International 10/20 Electroencephalography
(EEG) Coordinates—C3 (left hemisphere) and C4 (right hemisphere). However, considering
developmental reorganization, the optimal method for somatotopic localization has not been
established. Alternatively, Transcranial Magnetic Stimulation (TMS) with electromyographic recording
may determine location for maximal motor evoked potential (MEP) of target hand muscles. Here, we
compare two methods for localization of primary motor cortex (hand representation).
Methods: EEG C3/C4 coordinates and TMS locations were identified in 36 participants (23 typically
developing, 13 with hemiparesis), ages 8-21 (mean 13yrs, 9mo, SD 3yrs±1mo). Locations were recorded
using stereotactic neuronavigation (Brainsight) and distances were measured between C3/C4 and
TMS locations. Individual within-hemisphere comparisons were conducted.
Results: In typically developing participants, one TMS observation aligned with C3 (1/23 participants,
distance range: 0.7 – 5.24 cm), and none aligned with C4 (0/23 participants, range: 0.98 – 6.43 cm). In
participants with hemiparesis, no observations aligned within the lesioned hemisphere (0/11
participants, range: 0.83–3.9 cm) or the non-lesioned hemisphere (0/13 participants, range 0.29-13.92
cm). Two children with hemiparesis lacked an ipsilesional MEP.
Conclusions: Individual variability in brain somatotopic organization may influence surface scalp
localization of underlying primary motor cortex in children regardless of neurologic impairments. EEG
and TMS locations differed in all but one occurrence. Further investigation into location differences is
critical to consider when evaluating therapeutic benefits of tDCS interventions.
Poster Number: NM-169
Poster Name: Open-source System for Independent Control of Microelectrodes within the Brain
Authorship List: Luke Rosedahl, Matthew D. Johnson
Institution: University of Minnesota
Abstract:
Objective: Develop a system for manipulating microelectrodes during recording and stimulation that
provides visualization and logging of electrode depths/signals while being easily and cheaply built inlab.
Background: Commercial microdrive systems are often limited in the number of microelectrodes that
can be simultaneously manipulated and in the ability to adaptively control the depths of the
microelectrodes. Due to recent advances in 3D printing technology, it is now possible to create 3D
printed microdrive systems cheaply and quickly, providing opportunities to develop customized
microdrive systems that are open source and more user friendly.
Methods: Prior work in 3D printed microdrives provided a basic microdrive system that required
manual manipulation to advance the microelectrodes. Low error torsion cables for translating torque
while not inducing noise into the system were coupled with driving torque provided through Arduino
controlled stepper motors.
Results: The system is shown to provide 5um step resolution of microelectrode insertion and
retraction, which are controllable through an Arduino / MATLAB interface. This interface was then
coupled with a customized stereotactic neurosurgical navigation program known as Monkey Cicerone
to provide visualization and streamlined logging of all recording locations over time.
Conclusions: The resulting system provides a means to independently drive and interface with multiple
microelectrodes within the brain, while being cheaply buildable by research labs.
Future Directions: The system provides a comprehensive framework for performing multi-electrode
recording and stimulation experiments and a foundation to build more advanced technology including
adaptive neuronal tracking systems to adaptively compensate for brain micromotion and brain shift.
Poster Number: NM-170
Poster Name: Active Neuromodulation of Mesolimbic and Papez Circuit Connectivity
Authorship List: Erika Ross, Hoon-Ki Min, Joo Kim, Seongrok Han, Megan Settell, Charles Blaha,
Kendall Lee
Institution: Mayo Clinic
Abstract: Impaired memory retrieval in AD results from dysfunction within an integrated network,
including the medial temporal lobe, mammillary bodies, dorsomedial thalamus, posterior cingulate, and
connecting white matter tract of the fornix (Sperling et al., 2010). Recently, deep brain stimulation
(DBS) has been applied near the fornix to address the memory dysfunction associated with AD (Smith
et al., 2012). Based on these studies, we dissected the functional brain activation patterns from fornix
DBS, as measured by fMRI in a large animal model (pig). Here, we described significant blood oxygen
level dependent (BOLD) increase within many hippocampal circuit structures, as well as mesolimbic
dopaminergic structures including the nucleus accumbens (NAc). Additionally, we performed local
Glutamate (Glu) and dopamine (DA) receptor antagonism in the NAc. Here, we showed that Glu
receptor antagonism significantly decreased fornix DBS-elicited BOLD activation in the NAc, PFC, and
hippocampus, while DA antagonism significantly and selectively decreased BOLD activation in the NAc
and PFC. Hippocampal glutamatergic projections through the fornix to the NAc make close
appositional synapses with mesoaccumbens dopaminergic terminals on spiny GABAergic output
neurons in the NAc. The present findings suggest that fornix DBS modulates DA release via activation
of glutamatergic heteroreceptors on presynaptic dopaminergic terminals in the NAc. Combined with
our preliminary results showing that fornix DBS results in significant BOLD increase in the
hippocampus, entorhinal cortex, parahippocampal gyrus, PFC, and NAc, these data provide a useful
platform for developing translational studies investigating the effects of DBS on memory circuitry.
Poster Number: NM-171
Poster Name: Electrophysiological Correlates of Phosphenes Induced by Transcranial Alternating
Current Stimulation of the Visual Cortex
Authorship List: Abhrajeet Roy, Bryan Baxter, Bradley Edelman, Bin He
Institution: University of Minnesota
Abstract: Transcranial alternating current stimulation (tACS) is a promising method for noninvasively
modulating cortical activity through the administration of electric current across the scalp surface in a
frequency-specific manner. Furthermore, tACS over the occipital lobe can be used to induce
phosphenes: the perception of light in the absence of light entering the eye. However, the origins of
these tACS-induced phosphenes are not well understood, and it has been suggested that the induced
perception of light is due to retinal stimulation, versus direct activation of the visual cortex. The goal of
this study was to investigate the real time electrophysiological correlates of tACS-induced phosphenes
using high resolution electroencephalography (EEG). In a group of healthy human subjects, we
recorded 64-channel EEG during the administration of tACS over the visual cortex. Low-density
rubber tACS electrodes (5 cm x 5 cm) were placed on the scalp flanking the Oz EEG electrode
location and under the EEG cap, in order to target the occipital lobe. The perception of phosphenes
was induced using a 16 Hz bipolar stimulation waveform with an amplitude of +/-1000 uA. Importantly,
we found that tACS-induced changes in scalp electrical activity were localized over bilateral parietooccipital cortex and did not extend to frontal areas. These results suggest that the induced
phosphenes reflect a true activation of neurons in the visual cortex due to tACS. Further development
of these methods could lead to exciting new insights into the nature of phosphenes and more
generally, various neural correlates of perception and awareness.
Poster Number: NM-172
Poster Name: Assessing Neuromodulation In Vivo by Computerized Gait Analysis
Authorship List: Sina Safayi, Nick D. Jeffery, Sara K Shivapour, Mahdi Zamanighomi, Tyler J. Zylstra,
Joshua Bratsch-Prince, Saul Wilson, Chandan G. Reddy, Douglas C. Fredericks, John W. Miller,
George T. Gillies, Matthew A. Howard III
Institution: Iowa State University
Abstract:
Background: We are developing a novel intradural spinal cord (SC) stimulator designed to improve the
treatment of intractable pain and the sequelae of SC injury. In vivo ovine models of neuropathic pain
and moderate SC injury are being implemented for pre-clinical evaluations of this device, to be carried
out via gait analysis before and after induction of the relevant condition.
New Method: We extend previous studies on other quadrupeds to extract the three-dimensional
kinematics of the limbs over the gait cycle of sheep walking on a treadmill. Quantitative measures of
thoracic and pelvic limb movements were obtained from 17 animals. We calculated the total-error
values to define the analytical performance of our motion capture system for these kinematic
variables.
Results: Total-error averaged ~8%, ranging from 1% to 30% for individual gait variables. The post- vs.
pre-injury time delay between contralateral thoracic and pelvic-limb steps for normal and SC-injured
sheep increased by ~24 s over 100 steps. The pelvic limb hoof velocity during swing phase decreased,
while range of pelvic hoof elevation and distance between lateral pelvic hoof placements increased
after SC injury.
Comparison with Existing Method(s): The outcomes of our animal training protocols match those of
others, and the relative gait measurements reveal new kinematic findings with defined analytical
variations.
Conclusions: The kinematics measures in a single SC-injured sheep can be objectively defined as
changed from the corresponding pre-injury values, implying that this method can be used to assess the
impact of new neuromodulation strategies on the specific deficits exhibited by an individual.
Poster Number: NM-173
Poster Name: Neural Restoration via Loop-based Reinforcement in Parkinson's Disease: A Mechanism
of High Frequency Deep Brain Stimulation
Authorship List: Sabato Santaniello, Michelle M. McCarthy, Erwin B. Montgomery, John T. Gale,
Nancy Kopell, Sridevi V. Sarma
Institution: University of Connecticut
Abstract: We focus on two questions: (1) Why is DBS therapeutic only when the frequency of
stimulation belongs to a specific high range (130-180 Hz)? (2) What is the fundamental mechanism that
keeps high frequency DBS therapeutic even if the stimulation target is moved across brain structures?
We address these questions by developing a computational model of the direct pathway, including
motor cortex, thalamus, striatum, and GPi. Then, we used the model to study the effects of several
DBS settings via numerical simulations. Regarding to 1), we show that the therapeutic effects of DBS do
not entirely stem from local changes of the neuronal activity in the stimulation target but they also
depend in part on the fact that the motor loop is a closed reentrant system. Due to the closed-loop
nature, perturbations induced by DBS may travel along the system both forward and backward, and
overlap if the pulses are constantly spaced (DBS is regular) and close enough one to one another (DBS
is high frequency). This suggests that DBS impacts the entire loop, the therapeutic merit of clinicallyused DBS settings depends on the anatomy of the treated system, and DBS in different individuals
may require slightly different settings, which is consistent with clinical practice. Regarding to 2), we
show that the rendezvous occurs in the striatum and may determine a dominant discharge pattern that
percolates through the basal ganglia and restores the normal function of the thalamo-cortical subsystem, which is primarily involved in the selective dis-inhibition of motor commands.
Poster Number: NM-174
Poster Name: Inherent Variability of TMS Outcomes in a Single Subject
Authorship List: Rebekah L. Schmidt, Mo Chen, Teresa J. Kimberley
Institution: University of Minnesota
Abstract:
Background: Variable responses to non-invasive brain stimulation (NIBS) is a key area of concern to
the field of neuromodulation. The physiologic effects of NIBS are most frequently quantified with
transcranial magnetic stimulation (TMS) tests to determine changes in cortical excitability. Limited
literature exists regarding the inherent variance of TMS outcomes on a day-to-day basis which may
reduce the reliability of TMS outcomes to detect real changes in cortical excitability over multiple
sessions.
Methods: Healthy single-subject, multiple session (5 days) design to evaluate the variance TMS
outcomes: resting motor threshold (RMT), cortical silent period (CSP), short intracortical inhibition
(SICI), and intracortical facilitation (ICF). Changes from the first session were calculated and compared
to reported standard error of mean (SEM) and smallest real difference (SRD) values.
Results: RMT changes were nearly the same or less than estimated SEM (2.93) and SRD (7.48) values.
Mean CSP values were larger than SEM values (10.43ms) on three days and larger than SRD values
(28.91ms) on one day. SICI changes were larger than SRD (17%) and SEM (59%) values at some interstimulus intervals; however, there were specific intervals that were less variable and did not have
changes larger than SEM or SRD.
Conclusions: Healthy individuals experience day-to-day changes in cortical excitability that are larger
than estimated population SEM and SRD values reported in the literature. Detecting meaningful
physiologic effects from NIBS may require multiday averages to establish a steady baseline. Future
work should evaluate multiday variance in a larger sample and in various neurologic disorders.
Poster Number: NM-175
Poster Name: Ventral Tegmental Area (VTA) DBS Evoked Electrochemical and fMRI Responses
Authorship List: Megan Settell, Paola Testini, Joo Pyung Kim, Seongrok Han, Kevin Bennet, Kendall
Lee, Hoon-Ki Min
Institution: Mayo Clinic
Abstract: The ventral tegmental area (VTA) has been a region of interest in reward, addiction,
reinforcement, and learning as it provides the main dopaminergic outputs of the mesolimbic system.
Recently, stimulation of the VTA has been introduced in the treatment of psychiatric disorders such as
schizophrenia and depression. Dopaminergic projections from the VTA support it as a possible DBS
target for the treatment of depressive behavior (Friedman, Frankel et al. 2009). The complexity of the
anatomy of the VTA warranted the combined use of electrochemistry and functional imaging (fMRI) to
investigate the dopaminergic response induced by stimulation of this region in an animal model.
Adult swine were implanted with a stimulating DBS electrode targeting the VTA. Utilizing several
variations in stimulation parameters, the elucidated BOLD response was recorded and mapped across
all animals. Following fMRI, animals were implanted with carbon microfiber electrodes. These were
used in conjunction with the in-house built WINCS and MINCS system which can simultaneously
stimulate and take voltametric recordings of analytes, the release and concentration of which is
modulated by neural stimulation of the VTA. Recording electrode implantation targeted the brain
areas showing increased BOLD-signal. The sensing carbon fiber electrode was initially placed in the
caudate and progressively lowered while recording the dopamine signal. As the stimulation time and
amplitude increased, both the BOLD and extracellular dopamine signals increased. These increases
were evident in regions with higher BOLD signal increase, such as the nucleus accumbens and the
prefrontal cortex, as demonstrated during the depth study.
Poster Number: NM-176
Poster Name: Selective Sensing and Modulation of Brain Networks using High Density Intracranial
Electrode Arrays
Authorship List: Abbas Sohrabpour, Bin He
Institution: University of Minnesota
Abstract: Many brain disorders such as epilepsy or psychiatric disorders are network diseases with
distributed nodes over various brain regions with time varying activities. If multiple DBS leads are
placed close to regions of involvement, then using electrophysiological inverse techniques the
networks responsible for generating abnormal activity can be located. We hypothesize that using an
array of intracranial electrodes such as high density DBS leads and/or cortical electrodes, we can
obtain precise localization and imaging of pathological networks which can be distributed and varying
over time. We further hypothesize that using such array of intracranial electrodes including high
density DBS leads or/and cortical electrodes, we can offer enhanced targeted stimulation at critical
nodes in a subject specific and event specific manner.
We tested this hypothesis in a computer simulation study. DBS leads with 80 electrodes were placed
deep in the brain, bilaterally. Three functional nodes were simulated at thalamus, hippocampus and
cortex, respectively. Causal relations among the three nodes were assigned. Applying source imaging
techniques and dynamic causal modeling to the data, the three sources were localized with couple of
mm localization error and the causal nodes and the underlying network dynamics were identified (the
interactions and causal links between the nodes). After this relation was derived, the desired current
density to target these nodes and consequently interrupt pathological activity was achieved by
determining optimal current distribution of the DBS leads via solving the inverse problem. Target areas
were hit with good specificity. In summary, our simulation results suggest good specificity and
resolution in selectively sensing brain networks and targeting them using high density DBS leads.
Poster Number: NM-177
Poster Name: Estimating Source Extent from EEG/MEG by Means of a Sparse Signal Processing
Approach
Authorship List: Abbas Sohrabpour, Yunfeng Lu, Bin He
Institution: University of Minnesota
Abstract: We have proposed a new electric source imaging (ESI) that estimates the location and
extent of source distributions. Our proposed algorithm is based on the fact that detectable EEG/MEG
signals arise from brain regions that are spatially coherent and next to each other. This mathematically
translates to sparseness in gradient domain. Exploiting the sparsity of the underlying sources we have
been able to retrieve information about source extent. Another idea implemented in our algorithm is
to iteratively correct the estimation. This is achieved by penalizing locations that are associated with
current dipoles with lower amplitude, more than locations associated with higher amplitude (amplitude
of current dipoles are estimated at each iteration). This eliminates the need for subjective thresholding
of solutions.
In a Monte Carlo simulation where sources with different sizes where simulated at random locations
over the cortex with different amount of noise added to the scalp potentials, we were able to retrieve
extended sources. Our results show that the estimated extents are in good concordance with the true
extent and the average localization error is less than 5 mm with an average overlap of 80% between
estimated and simulated sources. While the estimated extent is not exact, i.e. there is some variance
when looking at true versus estimated extents in the Monte Carlo simulations, the results suggest an
unbiased estimate. In summary, our simulation results demonstrate that our proposed approach can
image source extent without ad hoc thresholding. This technique may have important applications in
brain functional mapping in general and imaging epileptogenic zone in particular.
Poster Number: NM-178
Poster Name: Manipulations of Visual Feedback Modulate Purkinje Cell Encoding of Task
Performance Consistent with a Forward Internal Model
Authorship List: Martha L. Streng, Laurentiu S. Popa, Timothy J. Ebner
Institution: University of Minnesota
Abstract: Computational theories postulate that effective motor control is achieved by a forward
internal model (FIM) that predicts the sensory consequences of motor commands. These internal
predictions are compared with the actual sensory consequences to generate sensory prediction errors
(SPEs), which are the error signals crucial for on-line movement control and motor learning. While
extensive evidence implicates the cerebellum in these computations, the mechanisms by which they
are encoded remain unknown. Simple spike (SS) discharge of Purkinje cells (PCs) has a dual
representation of performance errors and arm movement kinematics at lead and lag timing. The
predictive and feedback encoding has opposing effects on SS discharge, consistent with an SPE. In
support of this view, reducing visual feedback about errors selectively decreases feedback encoding.
Delaying visual feedback shifts the timing of predictive encoding to more feedforward times equal to
the duration of the delay, consistent with a FIM that has not adapted to the delay and makes
predictions with respect to the undelayed movement. Encoding of kinematics is unaffected, suggesting
it is driven primarily by arm movements.
Our results suggest that dual encoding of errors by SS discharge represents the predictive and
feedback signals necessary for the generation of SPEs. The differential effects of these manipulations
on error and kinematic encoding suggest the implementation of multiple FIMs. In this view, the
cerebellum processes predictions and feedback about the kinematics of arm movements and the taskrelevant performance errors to achieve optimal control.
Support: NIH R01 18338, NIH T32 GM008471, NSF IGERT DGE-1069104
Poster Number: NM-179
Poster Name: Cutaneous Sensory Nerve Function of the Scalp and Index Finger of Healthy Control
Subjects Classified by Fitzpatrick Skin Type
Authorship List: Leah Swanson, Elisabeth Hurliman, Kathleen Kane, Ana Lucia Junqueira, John
Connett, Maria Hordinsky
Institution: University of Minnesota
Abstract: Cutaneous nerve function has been found to be abnormal in many scalp diseases associated
with pain and itch. However, past studies of the scalp are limited in that there is minimal data on
healthy scalp innervation. In this study, we aimed to assess epidermal nerve fiber function of the
healthy scalp and extremity stratified by Fitzpatrick skin type as past studies of sensation across ethnic
groups have shown mixed results. 141 healthy control subjects underwent Current Perception
Threshold (CPT) testing using the NEUROMETER® CPT stimulator. The stimulator generates current
at 2000, 250, and 5 Hz to stimulate the A beta, A delta, and C fibers respectively. CPT testing was
completed on the index finger and occipital and temporal scalp. CPT values correspond to the
minimum electrical stimulus needed to conduct sensation. For each site and frequency combination,
difference in mean CPT was assessed with ANOVA and pairwise t tests with Bonferronni correction
for multiple comparisons. Results show that CPT values of the scalp were significantly lower than the
index finger. Between extremes of skin type, there were significant differences in CPTs for A beta and
A delta fibers of the scalp. For the index finger, there were no significant differences in CPTs across
skin types. This data provides previously unknown cutaneous nerve thresholds of the healthy scalp.
Unique scalp innervation is suggested by the greater sensitivity of the scalp compared to the index
finger. Additionally, these findings indicate that future studies of nerve function should stratify
subjects by skin type.
Poster Number: NM-180
Poster Name: Algorithms for Current Steering with Deep Brain Stimulation Arrays
Authorship List: Benjamin A. Teplitzky, Laura M. Zitella, Joe Xiao, Matthew D. Johnson
Institution: University of Minnesota
Abstract: Deep brain stimulation (DBS) therapy relies on accurate stimulus delivery to mm-scale brain
targets using mm-scale annular electrodes. DBS arrays with higher-density, circumferentiallydistributed electrodes have potential to improve patient outcomes by enabling more flexible current
steering and sculpting algorithms to compensate for sub-optimal lead placement. However, it is not
clear how many radial electrodes are necessary to effectively configure stimulation settings using
these arrays. Twenty-seven lead designs were evaluated using a combined finite element tissue
conductance model and multi-compartment axon model framework. Threshold-dependent regions of
activation (RoA) arising from monopolar and bipolar stimulation configurations were calculated using
the NEURON programing environment. RoAs were quantified using computer vision techniques and
deconstructed into feature-sets that were evaluated using machine learning classifiers. Single-cathode
monopolar stimulation shifted RoA center of mass 1-1.35mm in the direction of stimulation with
improved shifting resolution for leads with more radial electrodes. Shifting resolution was further
improved using non-uniform current distribution. Directional steering of the RoA related to electrode
angular separation and was not improved by multi-cathode or bipolar stimulation. Bipolar stimulation
resulted in anodic regions of activation that limited RoA sculpting. Compensation for DBS lead
misplacement up to 1.25 mm was possible for 1.5mm tall radial electrodes (n=2-7) while maintaining a
charge densities .150 C/cm2. Independent current sources greatly improved shifting resolution and the
use of bipolar stimulation for sculpting was found to be minimally advantageous. Accurate and
computationally efficient classification was achieved for monopolar stimulation with a k-nearest
neighbor classifier (error <5%) using a simple RoA derived feature-set.
Poster Number: NM-181
Poster Name: Computational Model of Stimulation-evoked Dopamine Release in a Porcine Model of
DBS
Authorship List: James K. Trevathan, Ali Yousefi, Peter J. Grahn, Aldo A. Mendez Ruiz, Andrea L.
McConico, J. Luis Lujan, Kendall H. Lee
Institution: Mayo Clinic
Abstract: Maximizing long-term efficacy of deep brain stimulation (DBS), a surgical therapy for
movement and psychiatric disorders, requires frequent empirical adjustment of stimulation parameters
to compensate for the dynamic and progressive nature of these disorders. Animal studies suggest that
therapeutic DBS is associated with specific patterns of neurotransmitter release. As such,
neurochemical monitoring coupled with feedback control of stimulation represents an excellent
mechanism for optimizing the fine balance between symptom reduction and stimulation-induced side
effects. Development of such system requires characterization of the non-linear and time-varying
neurotransmitter release evoked by DBS. In order to evoke dopamine release, the substantia nigra
pars compacta (SNc) / ventrotegmental area (VTA) region was stimulated with a randomized sequence
of parameters in a cohort of white farm swines. Evoked dopamine release was recorded using fast scan
cyclic voltammetry (FSCV) and a carbon fiber microelectrode placed within the striatum. The
relationship between stimulation parameters and evoked dopamine release was modeled using both
artificial neural network and Volterra series based approaches, chosen for their ability to represent
non-linear and time-variant systems. Results showed both modeling techniques were capable of
mathematically describing the forward and inverse non-linear and time-varying dynamics of stimulation
evoked neurotransmitter release. Development of these models provides an important first step
towards improving the current understanding of neurotransmitter release in the context of brain
stimulation. In turn, characterization of stimulation-evoked neurochemical release will be paramount to
the development of closed-loop DBS systems that can account and compensate for the non-linear
dynamics of neural circuits associated with therapeutic DBS.
Poster Number: NM-182
Poster Name: Total Variance Constrained Electrical Properties Tomography with a 16-channel
Transceiver Array Coil at 7T
Authorship List: Yicun Wang, Jiaen Liu, Xiaotong Zhang, Pierre-Francois Van de Moortele, Bin He
Institution: University of Minnesota
Abstract: Magnetic Resonance based Electrical Properties Tomography (MREPT) is a recently
introduced technique promising to provide quantitative electrical properties (EP) map of brain tissues
using Magnetic Resonance Imaging (MRI). Such information could be useful to guide precision
neuromodulation. Reconstruction of MREPT images relies on solving Maxwell’s Equations with
radiofrequency (RF) magnetic field distributions measured in MRI. In previous studies, Helmholtz
equation was utilized as the central equation, which assumes local homogeneity of EP, inducing severe
artifact on tissue boundaries. Other methods that considered EP gradient either suffered from
unsatisfactory accuracy or inconvenience of assigning integration seed points, presenting hurdles to
clinical practice. In this study, we have demonstrated that MREPT reconstruction with gradient
information could be achieved by solving an inverse problem to obtain the optimal solution that
minimizes squared residues regularized by Total Variance. Numerical simulation of RF field distribution
in a realistic human head model, as well as a phantom experiment using a 16-channel transceiver array
coil at 7T MRI, have been performed to evaluate the proposed algorithm. It has been shown that the
proposed algorithm is able to preserve tissue boundaries with excellent resolution and high accuracy.
Poster Number: NM-183
Poster Name: Acupuncture Analgesia on Awake Mice with Sickle Cell Disease
Authorship List: Ying Wang, Kalpna Gupta
Institution: University of Minnesota
Abstract:
Background: Acupuncture analgesia (AA) is widely used for pain management, which also effectively
alleviated associated pain of sickle cell disease (SCD) in patients. We aimed to develop
electroacupuncture (EA) treatment on awake mice to simulate clinical relevant condition on patients
as well as assessing the effectiveness of EA for chronic pain in SCD.
Methods: Male homozygous Berkeley mice (HbSS-BERK) that express exclusively (>99%) human sickle
hemoglobin as well as the littermate controls (HbAA-BERK) expressing normal human hemoglobin
were used.
EA treatment (once/3 d, frequency: 4 or 10 Hz, pulse wide: 100 μs, duration: 30 min) was performed on
bilateral acupoint GB30. Electrical stimulation was delivered by acupuncture needles connected with
an electrical stimulator. Sham-EA control was identically performed without electrical stimulation. Pain
level was evaluated by paw withdrawal frequency calculating the percentage of response among 10
stimulations onto the plantar surface.
Results: HbSS sickle mice showed varied analgesic effect with 43% positive responder, 43% moderate
responder and 14% non-responder by EA treatment.
Conclusions: EA on conscious free-moving mice simulates clinical condition on patients and excludes
the potential influence from restrainer or anesthetics. Variable analgesic response of sickle pain to EA
may suggest physiological individual difference to electrical stimulation. Combination of multiple
acupuncture-related techniques and acupoints for different clinical individuals is needed for better AA
evaluation in SCD.
Poster Number: NM-184
Poster Name: Repeated TDCS Administration as an Adjunct to Working Memory Training: Preliminary
Findings from a Randomized, Single-blind Control Trial
Authorship List: Yuanyuan Wang, Tasha M. Nienow, Angus W. MacDonald, Kelvin O. Lim
Institution: VA Healthcare System, University of Minnesota
Abstract:
Background: While cognitive remediation is currently the most effective method of enhancing
cognition in schizophrenia, results are modest, and there is considerable interest in identifying ways to
increase the effectiveness of this intervention. Transcranial direct current stimulation (tDCS) is a safe,
well-tolerated, non-invasive brain stimulation technique that enhances learning processes by
modulating synaptic strength. The aim of this proof of principle study was to examine whether
cognitive training was more efficacious when combined with tDCS.
Methods: To explore this hypothesis, working memory training was offered to a sample of 15
outpatients with schizophrenia who were randomized and blind to whether they received tDCS or
sham stimulation. Working memory training consisted of 48 1-hour sessions in which participants
completed adaptive, computer-based tasks that targeted working memory processes. Beginning in the
third week of the protocol, working memory training was augmented with 20 minutes of stimulation (1
mA tDCS/sham) twice a week. Participants received a total of 28 sessions of working memory training
combined with tDCS or sham stimulation. Efficacy of the intervention was assessed with performance
on trained (N-back) and untrained working memory tasks (MATRICS Consensus Cognitive Battery
Working Memory subtests).
Results: Ten participants completed all study procedures and drop-out appeared to be random. Postintervention, participants who received tDCS with working memory training performed at a higher
level (n = 6, d = 1.34) on a training task, a word N-back, than those who received sham (n = 4, d = .81). In
addition, we found evidence that administration of tDCS promoted greater generalization of cognitive
training. Recipients of tDCS also demonstrated better performance on a novel version of a training
task (picture N-back) than sham participants (d = 1.21 vs .63). In addition, participants in the tDCS
condition demonstrated greater transfer to untrained working memory tasks than those in the sham
group: MATRICS Spatial Span (d = .64 vs .36) and Letter Number Span (d = .36 vs .0).
Conclusions: While preliminary, these results suggest that pairing tDCS with working memory training
produces a more effective learning experience. Training gains were more robust, and there was
greater transfer to untrained tasks when cognitive training was paired with tDCS, suggesting a
complementary procedure to increase the efficacy of cognitive remediation.
Poster Number: NM-185
Poster Name: Analysis of Computational Models of Deep Brain Stimulation using Spherical Statistics
Authorship List: YiZi Xiao, Matthew D. Johnson
Institution: University of Minnesota
Abstract: Computational field and neuron models of deep brain stimulation (DBS) have played a key
role in investigating the mechanisms of action of DBS therapies. By estimating a volume of tissue
directly modulated by DBS, one can relate the pathways within those volumes to the therapeutic
efficacy of a particular DBS setting. In this study, we applied the tools of spherical statistics to quantify
morphologies of modulated volumes in a computational model of DBS. We investigated stimulation
through a novel 32 electrode DBS array on a population of 5000 thalamocortical relay neurons. We
developed neuronal distribution hypotheses, tested them for goodness of fit, and estimated model
parameters for fitted distributions. We also analyzed the change in statistical model parameters by: 1)
increasing current amplitude and 2) fixing the current amplitude and analyzing the distribution of
activated neurons in concentric spherical shells. We found that existing spherical statistical models are
well suited to characterize distributions of neurons activated by DBS. In addition, changes in data
selection (soma/axon nodes) and spherical origin can critically change how results are interpreted.
Increasing current amplitude activated larger populations of neurons with decreased concentration
about the mean direction. However, when the current amplitude was constant, activated neurons in
outer spherical shells were more concentrated about their mean direction. We believe this spherical
statistical framework can be a useful tool for studies that require objectively quantifying the spatial
distribution of neuronal activity in 3D, such as with in-vivo electrophysiological recordings, in-vivo
calcium imaging, and histological labeling of neurons.
Poster Number: NM-186
Poster Name: Network Analysis of Intrinsic Functional Brain Connectivity in Borderline Personality
Disorder
Authorship List: Tingting Xu, Kathryn R. Cullen, Kelvin O. Lim, S. Charles Schulz, Keshab K. Parhi
Institution: University of Minnesota
Abstract:
Background: Borderline personality disorder (BPD) is a severe mental disorder affecting about 1.3 % of
the general population. Neuroimaging studies have revealed both structural and functional
abnormalities in specific brain regions and connections in BPD. However little is known about the
alternations in topological organizations of whole brain networks in BPD patients.
Methods: Resting-state functional magnetic resonance imaging (fMRI) data were acquired from twentyone BPD patients and ten healthy control subjects. Wavelet analysis was performed to compute
frequency-dependent correlation matrices of 90 brain regions; and the functional brain networks were
constructed by thresholding the correlation matrices at various graph density. The topological
properties of the brain networks, e.g., small-worldness, efficiency, clustering, path length and centrality,
were obtained using graph theory-based complex network analysis.
Results: Both the BPD patients and the healthy controls show small-world architecture in functional
brain networks, suggesting a balance between functional segregation and integration. However,
compared with healthy subjects, BPD patients show altered quantitative values in the global
properties characterized by decreased path length, increased size of largest connected components,
local efficiency and global efficiency, implying a shift toward randomization in their brain networks.
Conclusions and Significance: The findings in this study suggest disrupted topological organizations of
functional brain networks in BPD. The discriminating network measurements may be useful as imagingbased biomarkers for BPD identification.
Poster Number: NM-187
Poster Name: Ultrasound-induced Directional Lorentz-current Stimulation: A Phantom Study
Authorship List: Kai Yu, Dalong Liu, Emad S. Ebbini, Bin He
Institution: University of Minnesota
Abstract: To manage the brain conditions, the electrical-based approaches, such as deep brain
stimulation, transcranial magnetic stimulation, have been effectively employed in treating diseased
neurological state. Further, a light-based method, optogenetic neuromodulation has also demonstrated
its superior spatio-temporal specificity in mapping brain circuits and controlling motor behavior.
However, as these approaches are invasive, to identify a non-invasive but efficient strategy for
modulating neural activity thus becomes one of the current challenges in neuroscience. Hence,
ultrasound technique has attracted great attention and been used to mediate the stimulation process.
In our proposed non-invasive stimulation scheme, a focused ultrasound beam is deployed to vibrate a
local volume of a conductive object with the ultrasound radiation pressure. With the existence of an
external magnetic field, a corresponding Lorentz current is thus induced at the focal spot of the
ultrasound beam. When the orientation of the ultrasound projection is changing, the direction of the
induced electric current is thus altered accordingly. In this present study, we have tested this
ultrasound-induced directional Lorentz-current stimulation in saline gel phantoms with dual mode
ultrasound array. To receive the ultrasound radiation pressure, a certain concentration of cellulose is
introduced as scatterer in the preparation of the phantom. To determine the induced current, a pair of
unipolar needle electrodes and electric potential measurement are used. The experimental results
indicate that this non-invasive stimulation possesses millimeter-level spatial specificity, and as this
approach is closely coupled with ultrasound, this stimulation strategy would provide both mechanical
and electrical neural modulation effect.
Poster Number: NM-188
Poster Name: Thalamocortical Relationship in Epileptic Patients with Generalized Spike and Wave
Discharges
Authorship List: Huishi Zhang, Zhiyi Sha, John Mundahl, Sa Liu, Yunfeng Lu, Thomas R. Henry, Bin He
Institution: University of Minnesota
Abstract: Epilepsy is one of the most prevalent neurological diseases. There are two broad categories
of epilepsies, focal or partial epilepsy, which has a confined range of influence, and idiopathic
generalized epilepsy (IGE), where the whole or a larger portion of the brain are affected without
apparent cause. Therefore, for effective treatment, it is important to understand the underlying
network which generates IGE activity and through which epileptic activity propagates. The aim of the
present study was to study the thalamocortical relationship using three non-invasive imaging
modalities in a group of IGE patients. We specifically investigated the roles of the mediodorsal nuclei
in the thalami and the medial frontal cortex in generating and spreading IGE activities. We
hypothesized that the connectivity between these two structures is key in understanding the
generation and propagation of epileptic activity in brains affected by IGE. Using three imaging
modalities of EEG, fMRI and EEG-informed fMRI, we identified important players in generation and
propagation of generalized spike-and-wave discharges (GSWDs). Medial frontal lobe and the
mediodorsal nuclei appeared to serve important roles in orchestrating GSWDs in the thalamus and the
cortex. Thalamus, especially the mediodorsal nuclei, may serve as potential targets for deep brain
stimulation to provide more effective treatment options for patients with drug-resistant generalized
epilepsy.
Poster Number: NM-189
Poster Name: Fitting Models of Parkinson’s Disease Finger-tapping Dynamics Using a Particle Filter
Authorship List: Simeng Zhang, Matthew D. Johnson, Scott E. Cooper
Institution: University of Minnesota
Abstract: Patients with Parkinson's disease exhibit characteristic abnormalities of rhythmical
movements known as "hastening" or "festination" in which frequency is pathologically increased.
Rhythmic finger-tapping (MDS-UPDRS item 3.4) is used clinically to quantify these abnormalities. A
successful model of the dynamics of Parkinsonian rhythmical movements could serve as the basis for
closed loop deep brain stimulation for treatment-refractory symptoms affecting rhythmical activities of
daily living such as speech and gait.
We constructed two models of Parkinson's finger-tapping dynamics, one with a single stable limit cycle
attractor (single-basin model), and the other with two stable limit cycle attractors, representing normal
and festinating movements (double-basin model). In 29 Parkinson's patients and 20 control patients
performing finger-tapping, we measured motion of the proximal thumb and index finger with angular
velocity transducers. We fit the models to the data using a particle filter, assessing goodness of fit by
cross-validation. We are now comparing goodness-of-fit between the single- and double-basin models
in Parkinson's and control subjects, planning to assess the the filter for detecting transitions between
normal and festinating regimes.
Poster Number: NM-191
Poster Name: Orexin Neuron Stimulation in Mice Protects Against High Fat Diet-induced Weight Gain
Authorship List: Anastasia Zink, Charles Billington, Catherine Kotz
Institution: University of Minnesota
Abstract: Sedentary lifestyles are a major risk factor for metabolic disease and negatively influence
overall health. Low levels of the neuropeptide, orexin, are associated with reduced activity levels and
obesity in both humans and animals. In this study, pharmacosynthetic neuromodulation (Designer
Receptors Exclusively Activated by Designer Drugs; DREADD) was used to test if stimulation of
orexin-producing neurons in the lateral hypothalamic area increases spontaneous physical activity and
promotes weight loss under obesogenic conditions. Transgenic mice expressing the DNA-recombinase,
Cre, in orexin neurons received bilateral virus injections into the lateral hypothalamic area. The virus
contained a Cre-dependent construct encoding the excitatory DREADD, hM3Dq, a modified Gqcoupled GPCR. After recovery, systemic injections of the designer drug, Clozapine-N-Oxide (CNO),
were given to selectively activate DREADD-containing orexin neurons. CNO is a highly-soluble
DREADD ligand that is biologically inert in rodents. An acute dose of CNO (5mg/kg; IP) increased time
spent moving at least three-fold in the two hours post-injection (p<0.001). Adult male mice were then
housed on high fat diet and given CNO or control vehicle (saline), once daily, for five consecutive days.
CNO treatments yielded healthy body weights that were not significantly different from standard
chow diet conditions; whereas, animals that received vehicle gained significantly more weight (p<0.001)
and adiposity (p=0.02). This study indicates the orexin neuropeptide system is a potent therapeutic
target for increasing non-exercise-dependent physical activity in the treatment of obesity and related
diseases.
Poster Number: NM-192
Poster Name: Model-based Parameter Sweeps of Pedunculopontine Nucleus Deep Brain Stimulation
Authorship List: Laura M. Zitella, Benjamin A. Teplitzky, Heather M. Hudson, Katelynn Brintz, Yuval
Duchin, Noam Harel, Kenneth B. Baker, Jerrold L. Vitek, Matthew D. Johnson
Institution: University of Minnesota
Abstract: The pedunculopontine nucleus (PPN), a component of the mesencephalic locomotor region,
is an investigational deep brain stimulation (DBS) target for treatment of freezing of gait in Parkinson’s
disease. To further investigate the pathways modulated during PPN-DBS, we developed a subjectspecific finite element model coupled with multi-compartment neuron models of PPN-DBS. The
models were fit to the anatomy (pre-operative high-field 7T MRI) and DBS lead position (postoperative CT) of a nonhuman primate that was implanted in the PPN area with a DBS lead. Postmortem histology was used to confirm these locations. In this study, we 1) investigated the effects of
adding tissue conductance anisotropy to the brainstem model and 2) performed model parameter
sweeps to determine model sensitivity. A nonlinear warping algorithm and probabilistic diffusion
tractography were used to delineate pathways associated with the PPN, superior cerebellar peduncle,
oculomotor nerve, and central tegmental tract. This anatomical framework was coupled with a finite
element model simulating the voltage distribution in the brainstem during DBS. For the anisotropic
model, conductivity values were calculated from diffusion tensors and interpolated onto the mesh,
while the isotropic model conductivity was set to 0.3 S/m. The models were evaluated with varying
axon diameter, lead location, and conductivity values. The results showed that lead location was an
important factor, but accurate axon diameter was even more critical. Furthermore, the inclusion of
anisotropy in the models was critical in that the anisotropic models resulted in a much lower threshold
for activation, correlating much better with in vivo behavior results.