Search

Data to reproduce statistical analysis and Figure corresponding to: https://doi.org/10.1101/2022.06.20.496467

Humans and other primates explore visual scenes by active sensing , using saccadic eye movements to relocate the fovea and sample different bits of information multiple times per second. Recent reports indicate that saccades induce a phase reset of ongoing neuronal oscillations in a distributed network including primary and higher-order visual cortices and medial temporal lobe. As a result, neuron ensembles in these regions are shifted to a momentary high excitability state at the time visual input enters the system; i.e., just after fixation; among other things, this would amplify the neuronal response to visual input. The extent of the brain’s circuitry modulated by saccades is not yet known. Here, we evaluate the possibility that saccadic phase reset impacts the anterior nuclei of the thalamus (ANT) using field potential recordings in the ANT in three surgical patients undergoing deep brain stimulation for treatment of epilepsy. Given that ANT are strongly connected to frontal/prefrontal cortex, hippocampus and cingulate cortex - brain networks that support active sensing, we hypothesized that ANT would display similar saccade-related modulation of excitability. In all three patients, we found saccade-related phase concentration, peaking at 3-4 Hz, coincident with suppression of Broadband High frequency Activity (BHA; 70-150Hz), a signal that correlates with local neuronal spiking activity. Our results provide evidence for saccade-related modulation of neuronal excitability dynamics in the ANT, consistent with the idea that these nuclei are part of the brain circuitry underlying visual active sensing.

Broadband High-frequency Activity (BHA; 70-150 Hz), also known as "high gamma," a key analytic signal in human intracranial recordings is often assumed to reflect local neural firing (multiunit activity; MUA). Accordingly, BHA has been used to study neuronal population responses in auditory (1,2), visual (3,4), language (5), mnemonic processes (6-9) and cognitive control (10,11). BHA is arguably the electrophysiological measure best correlated with the Blood Oxygenation Level Dependent (BOLD) signal in fMRI (12-13). However, beyond the fact that BHA correlates with neuronal spiking (12, 14-16), the neuronal populations and physiological processes generating BHA are not precisely defined. Although critical for interpreting intracranial signals in human and non-human primates, the precise physiology of BHA remains unknown. Here, we show that BHA dissociates from MUA in primary visual and auditory cortex. Using laminar multielectrode data in monkeys, we found a bimodal distribution of stimulus-evoked BHA across depth of a cortical column: an early-deep, followed by a later-superficial layer response. Only, the early-deep layer BHA had a clear local MUA correlate, while the more prominent superficial layer BHA had a weak or undetectable MUA correlate. In many cases, particularly in V1 (70%), supragranular sites showed strong BHA in lieu of any detectable increase in MUA. Due to volume conduction, BHA from both the early-deep and the later-supragranular generators contribute to the field potential at the pial surface, though the contribution may be weighted towards the late-supragranular BHA. Our results demonstrate that the strongest generators of BHA are in the superficial cortical layers and show that the origins of BHA include a mixture of the neuronal action potential firing and dendritic processes separable from this firing. It is likely that the typically-recorded BHA signal emphasizes the latter processes to a greater extent than previously recognized.