Seyhmus Guler




Electrical and Computer Engineering
Northeastern University

409 Dana Research Center
Northeastern University
360 Huntington Ave
Boston, MA 02115










I received my B.Sc. degree in Electrical and Electronics Engineering from Bilkent University, Ankara, Turkey.
I am currently a Ph.D. student at Northeastern University, working with Dana Brooks.

Research Interests:

Optimizing stimulus pattern for brain stimulation: Electrical stimulation of the brain to modulate on-going brain activity has become a topic of great interest in both clinical and research communities. One of brain stimulation techniques, transcranial current stimulation (tCS), makes use of scalp electrodes to inject low amplitude electrical currents to modulate a particular brain region of interest (ROI). In order to more precisely control the flow of current injected into the head, there has been increasing interest in the use of arrays of many small electrodes, known as “dense electrode arrays”. However, there is need for systematic optimization methods to calculate electrode current patterns of such arrays that satisfy both a meaningful optimality criterion, and a variety of safety and practicality constraints. In this project, we formulate and solve an efficient method that optimizes electrode stimulus patterns for dense array tCS while respecting thoughtfully chosen safety constraints. Specifically, the proposed method provides a globally optimal and unique electrode stimulus pattern that maximizes the current along a predefined desired direction in the ROI and satisfies three safety constraints imposed on the total injected current, on individual electrode currents, and on the current power in the brain outside the ROI. The proposed method has the capability of imposing additional power constraints, is compatible with spatially extended cortical ROIs, and can be extended to address stimulus pattern optimization regarding other brain stimulation modalities.

Seeking practical stimulus patterns: In dense array tCS or ECoG stimulation, each electrode is assumed to have its own designated current source to allow independently controlled current values from this electrode. This, however, may be impractical and rather expensive and thus we adapt our optimization problem to find stimulus patterns that use only a few current sources to feed all the electrodes in the array. This requires us to solve a combinatorial optimization problem, which we tackle by using branch and bound algorithm.

Branch and bound in action