Locally induced traveling waves generate globally observable traveling waves.

Abstract

Cortical traveling waves have been proposed as a fundamental mechanism for neural communication and computation. Methodological uncertainties currently limit the interpretability of non-invasive, extracranial traveling wave data, sparking debates about their cortical origin. Studies using EEG or MEG typically report waves that cover large portions of the sensor array which are often interpreted as reflecting long range cortical waves. Meanwhile, invasive, intracranial recordings in humans and animals routinely find both local, mesoscopic waves and large scale, macroscopic waves in cortex. Whether the global sensor-array waves found with EEG/MEG necessarily correspond to macroscopic cortical waves or whether they are merely projections of local dynamics remains unclear. In this study, we made use of the well-established retinotopic organization of early visual cortex to generate traveling waves with known properties in human participants (N=19, 10 female, 9 male) via targeted visual stimulation, while simultaneously recording MEG and EEG. The inducer stimuli were designed to elicit waves whose traveling direction in mesoscopic retinotopic visual areas depends on stimulus direction, while leaving macroscopic activation patterns along the visual hierarchy largely unchanged. We observed that the preferred direction of traveling waves across the sensor array was influenced by that of the visual stimulus, but only at the stimulation frequency. Comparison between single-trial and trial-averaged responses further showed considerable temporal variation in traveling wave patterns across trials. Our results highlight that under tight experimental control, non-invasive, extracranial recordings can recover mesoscopic traveling wave activity, thus making them viable tools for the investigation of spatially constrained wave dynamics.Significance statement Non-invasively obtained time-resolved neuroimaging data is often thought to primarily reflect neural dynamics on the largest spatial scales. In the context of cortical traveling waves, this assumption can lead to a misinterpretation of spatio-temporal patterns observed in the sensor array. We here show that it is in principle possible that the global sensor array data is dominated by spatially constrained, local cortical traveling wave activity. Our findings crucially inform the ongoing discussion about the origin of traveling waves observed in surface recordings.