Neural Synchrony Amongst Computational Units of the Language System
Overview: In this project, I used ICE to record signals from populations of neurons throughout the left hemisphere in a dozen patients’ brains, and assessed when and to what degree different populations synchronized their electrical activity (e.g., fired in phase with each other).
Neural synchronization is a phenomenon that may be, or be related to, a fundamental way that the brain communicates information among its many computational entities. This project is part of my larger goal to help understand how the human brain coordinates information processing, in order to capitalize on both its highly specialized ‘local’ resources and its broad networks that bind together multiple elementary faculties into high-order faculties like language.
The measure of neural synchrony employed here was “phase-locking” (see Lachaux, et al., Hum. Brain Mapping 8, 194 (1999)). The degree of phase-locking between two signals should be independent of their relative magnitude; and can be measured irrespective of the phase delay between the signals. A high phase-locking value indicates a high degree of consistency in the relative phase (difference in phase angle) between two recording sites, across a set of epochs such as task trials.
This project has produced several sets of results. First of all, when I focused on known and canonical functional-anatomical divisions of the brain (such as Broca’s area, Angular Gyrus, V1/V2, Wernicke’s area, Premotor cortex, etc), I found interesting dynamics of putative transient functional connectivity across pairings of these divisions that the literature would predict to work together (though with some surprises). These interactions will feature in an upcoming manuscript. The focus will be on the fine-grained timing of these interactions; and some indications of their function in information space (though there can only be limited evidence of such, given our state of understanding of what ‘information’ means in the brain). Also, these large functional-anatomical divisions will be respected, in part due to their precedent in the relevant literature, even though neural computations underlying cognition may well be best described at a finer spatial scale or indeed along more dimensions than just the spatial.
Another set of results from the phase-locking analysis converged toward a separate large-order pattern I have observed in the data from several patients. This pattern is very exciting to me and has prompted me to think at length, actually for several years, about what challenges the human brain needs to solve in order to produce high-level cognition from a collection of hyperspecialized computational units. I am currently writing up the results and my meta-level interpretation into a manuscript I hope may become my most useful paper or research product to date. Fingers crossed! Once I have received further feedback on the work, I will outline it here.
Detailed descriptions coming soon….