Poster No:
35
Submission Type:
Abstract Submission
Authors:
Uma Mohan1, John Wittig2, Oceane Fruchet2, Sara Inati2, Kareem Zaghloul2
Institutions:
1NIH, Bethesda, MD, 2NIH, NINDS, Bethesda, MD
First Author:
Co-Author(s):
Introduction:
Direct electrical brain stimulation combined with intracranial electrophysiological recordings hold the potential to modulate and test the functional role of neural activity in the awake human brain. While clinicians have used direct electrical brain stimulation for functional mapping and treatment of neurological and psychiatric disorders, the effects of stimulation on neural activity are poorly understood. Changes in neural activity from stimulation in local and remote areas are often highly complex and variable. Stimulation has most often been delivered at locations individually, however, simultaneous or patterned stimulation at multiple locations holds the potential to modulate distributed networks more precisely.
Methods:
To better understand and precisely control the responses to stimulation in individual patients, we first took the approach of modelling the effects of stimulation on neural dynamics across the brain. We collected human electrocorticographic recordings from 8 neurosurgical epilepsy patients while systematically delivering cortical stimulation at different frequencies, amplitudes, durations, and locations while patients were at rest. Using a dynamic linear state-space model framework, we fit input-output models to timecourses of neural activity, represented by high frequency activity, while patients received stimulation.
Results:
We first show that dynamic responses in brainwide neural activity following stimulation at individual locations across brain regions can be accurately predicted using latent state space models. We further show patient-specific models build can be used to predict responses to novel stimulation locations. Lastly, we analyzed changes in large-scale neural activity in response to multisite stimulation and compare these responses to those predicted from patient-specific state-space models built while patients were stimulated at individual locations. We found that we able to reliably predict the timecourse of responses to stimulation delivered at novel combinations of multiple locations.
Conclusions:
The ability to characterize and model neural responses to novel locations as well as patterns of multisite stimulation could allow clinicians and researchers to design stimulation protocols for precise modulation of neural activity. Stimulation parameters and patterns may be selected to elicit specific changes to ongoing behaviorally relevant neural signals in the human brain to modulate higher-order cognitive functions and to more effectively probe functional brain networks and treat neurological disorders.
Brain Stimulation:
Deep Brain Stimulation
Direct Electrical/Optogenetic Stimulation 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
EEG/MEG Modeling and Analysis
Exploratory Modeling and Artifact Removal
Keywords:
Cognition
Cortex
ELECTROCORTICOGRAPHY
ELECTROPHYSIOLOGY
Memory
Single unit recording
Tractography
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Workflows
1|2Indicates the priority used for review
Provide references using author date format
Yang, Y. (2021). "Modelling and prediction of the dynamic responses of large-scale brain networks during direct electrical stimulation." Nature biomedical engineering, 5(4), 324-345.