Skip to content

Direct Electrical Stimulation Networks Reflects Structural and Functional Connectivity in the Human Brain

Key findings

  • Direct electrical stimulation, particularly single-pulse electrical stimulation, has been proposed as a tool to understand connectivity in the brain, but there are debates about whether it represents functional, effective or structural connectivity
  • Working with 11 patients with epilepsy undergoing intracranial recordings, researchers at Massachusetts General Hospital measured stimulation-induced connectivity and compared it with resting-state functional, effective and structural connectivity
  • Structural connectivity measures produced very similar networks to those derived by single-pulse electrical stimulation; however, this similarity was driven in large part by the underlying relationship between distance and connectivity
  • Different connectivity measures, including those derived from active stimulation-based probing, seem to measure different, complementary aspects of regional relationships in the brain
  • When channels were segmented into local versus distant, structural connectivity was similar to distant cortico-cortical evoked potentials (CCEP) spread, while local CCEP networks reflected functional connectivity measures

Measurement of brain connectivity is a standard strategy for quantifying brain networks and the coupling between brain areas. Connectivity measures derived from cortico-cortical evoked potentials (CCEPs), which are elicited by direct intracranial electrical stimulation, have many advantages over other tools for estimating brain networks. For example, it can estimate the connectivity of multiple cortical areas and reveal changes related to underlying pathology.

However, it has been unclear how the CCEP network relates to the conventional types of brain network connectivity (functional, effective and structural), which can be determined without external stimuli.

Angelique C. Paulk, PhD, instructor in Neurology, and Sydney Cash, MD, PhD, co-director of the Center for Neurotechnology and Neurorecovery (CNTR) of the Department of Neurology at Massachusetts General Hospital, and colleagues have found evidence that CCEPs are a reliable measure of brain connectivity. However, as they explain in NeuroImage, the degree to which other connectivity measures correlated with the CCEP depended on distance and the type of measure.

Study Methods

The researchers recorded CCEPs induced by single-pulse electrical stimulation at 104 sites across 11 participants with intractable epilepsy (2–23 stimulation sites per patient). The patients were being implanted with bilateral intracranial electrodes to identify seizure onset zones.

The voltage response spread of CCEPs across sites was compared with six measures of connectivity during resting-state conditions:

  • Functional connectivity, which identifies channels or brain regions with correlated activity (three measures)
  • Effective connectivity, which identifies whether channels or brain regions may causally influence one another (two measures)
  • Structural connectivity was estimated using diffusion tensor imaging (DTI), which measures the anisotropy of water molecules, allowing partial reconstruction of white matter tracts in the brain

Key Findings

  • Across the 11 participants, 53 stimulation sites produced at least one above-threshold response
  • Of the eight participants who had DTI scans, 34 of 68 stimulation sites produced at least one above-threshold response
  • In estimates of similarity not controlled for distance from the recording site, structural connectivity based on diffusion tractography imaging was the strongest correlate of connectivity represented by the CCEP network
  • However, controlling for distance abolished the advantage of structural connectivity above functional and effective estimation methods
  • When channels were segmented into local versus distant, structural connectivity was the most similar to distant CCEP spread, even after controlling for distance, while local CCEP networks reflected functional connectivity measures

Interpreting the Findings

The relationship between distance and connectivity is probably attributable to a variety of factors, including volume conduction influencing measured activity and the underlying tendency of brain regions to form both physical and functional connectivity with neighboring areas. Still, distance-corrected connectivity was better suited for determining which non-invasive methods of connectivity estimation predict the effects of direct electrical stimulation.

These results suggest that different connectivity measures, including those derived from active stimulation-based probing, measure different, complementary aspects of regional relationships in the brain, although functional and effective methods seem particularly well-suited for predicting local stimulation effects. This relationship could be crucial when designing stimulation paradigms for clinical and research purposes.

Learn more about the Center for Neurotechnology and Neurorecovery

Learn more about research in the Department of Neurology

Related topics

Related

Physicians at Massachusetts General Hospital have demonstrated the novel, effective use of responsive neurostimulation for treating status epilepticus in a patient who did not respond to antiepileptic drugs, immunomodulatory therapy, ketogenic diet or transcranial magnetic stimulation.

Related

Jimmy C. Yang, MD, and Sydney S. Cash, MD, PhD, of the Department of Neurology, and colleagues made microelectrode recordings during neurosurgeries and identified cortex microdomains that seem to contribute to the seizure network—an important step toward understanding how to develop better treatments for epilepsy.