High density single-unit human cortical recordings using the Neuropixels probe

A fundamental unit of neural computation is the action potential. While significant advances have been made in the ability to sample action potentials of large numbers of individual neurons in animal models, translation of these methodologies to humans has been lacking due to clinical time constraints, electrical noise in the operating room, and reliability of the methodology. Here we present a reliable method for intraoperative recording of dozens of neurons in humans using the Neuropixels probe, yielding up to ~100 simultaneously-recorded single-units (n=596 across 11 recordings in 8 participants). Most single-units were active within 1 minute of reaching target depth, compatible with clinical time constraints. Cell pairs active close in time were spatially closer in most recordings, demonstrating the power to resolve complex cortical dynamics. Altogether, this approach provides access to population single-unit activity across the depth of human neocortex at scales previously only accessible in animal models.

Sensory cortical dynamics during optical microstimulation training

Primary sensory cortex is a key locus of plasticity during learning. Exposure to novel stimuli often alters cortical activity, but isolating cortex-specific dynamics is challenging due to extensive pre-cortical processing. Here, we employ optical microstimulation of pyramidal neurons in layer (L) 2/3 of mouse primary vibrissal somatosensory cortex (vS1) to study cortical dynamics as mice learn to discriminate microstimulation intensity. Tracking activity over weeks using two-photon calcium imaging, we observe a rapid sparsification of the photoresponsive population, with the most responsive neurons exhibiting the largest declines in responsiveness. Following sparsification, the photoresponsive population attains a stable rate of neuronal turnover. At the same time, the photoresponsive population increasingly overlaps with populations encoding whisker movement and touch. Finally, we find that mice with larger declines in responsiveness learn the task more slowly than mice with smaller declines. Our results reveal that microstimulation-evoked cortical activity undergoes extensive reorganization during task learning and that the dynamics of this reorganization impact perception.

Delirium, Caffeine, and Perioperative Cortical Dynamics

Delirium is a major public health issue associated with considerable morbidity and mortality, particularly after surgery. While the neurobiology of delirium remains incompletely understood, emerging evidence suggests that cognition requires close proximity to a system state called criticality, which reflects a point of dynamic instability that allows for flexible access to a wide range of brain states. Deviations from criticality are associated with neurocognitive disorders, though the relationship between criticality and delirium has not been formally tested. This study tested the primary hypothesis that delirium in the postanesthesia care unit would be associated with deviations from criticality, based on surrogate electroencephalographic measures. As a secondary objective, the impact of caffeine was also tested on delirium incidence and criticality. To address these aims, we conducted a secondary analysis of a randomized clinical trial that tested the effects of intraoperative caffeine on postoperative recovery in adults undergoing major surgery. In this substudy, whole-scalp (16-channel) electroencephalographic data were analyzed from a subset of trial participants (n = 55) to determine whether surrogate measures of neural criticality – (1) autocorrelation function of global alpha oscillations and (2) topography of phase relationships via phase lag entropy – were associated with delirium. These measures were analyzed in participants experiencing delirium in the postanesthesia care unit (compared to those without delirium) and in participants randomized to caffeine compared to placebo. Results demonstrated that autocorrelation function in the alpha band was significantly reduced in delirious participants, which is important given that alpha rhythms are postulated to play a vital role in consciousness. Moreover, participants randomized to caffeine demonstrated increased alpha autocorrelation function concurrent with reduced delirium incidence. Lastly, the anterior-posterior topography of phase relationships appeared most preserved in non-delirious participants and in those receiving caffeine. These data suggest that early postoperative delirium may reflect deviations from neural criticality, and caffeine may reduce delirium risk by shifting cortical dynamics toward criticality.

Bridging physiological and perceptual views of autism by means of sampling-based Bayesian inference

Theories for autism spectrum disorder (ASD) have been formulated at different levels: ranging from physiological observations to perceptual and behavioral descriptions. Understanding the physiological underpinnings of perceptual traits in ASD remains a significant challenge in the field. Here we show how a recurrent neural circuit model which was optimized to perform sampling-based inference and displays characteristic features of cortical dynamics can help bridge this gap. The model was able to establish a mechanistic link between two descriptive levels for ASD: a physiological level, in terms of inhibitory dysfunction, neural variability and oscillations, and a perceptual level, in terms of hypopriors in Bayesian computations. We took two parallel paths: inducing hypopriors in the probabilistic model, and an inhibitory dysfunction in the network model, which lead to consistent results in terms of the represented posteriors, providing support for the view that both descriptions might constitute two sides of the same coin.

The hippocampus as the switchboard between perception and memory

Adaptive memory recall requires a rapid and flexible switch from external perceptual reminders to internal mnemonic representations. However, owing to the limited temporal or spatial resolution of brain imaging modalities used in isolation, the hippocampal–cortical dynamics supporting this process remain unknown. We thus employed an object-scene cued recall paradigm across two studies, including intracranial electroencephalography (iEEG) and high-density scalp EEG. First, a sustained increase in hippocampal high gamma power (55 to 110 Hz) emerged 500 ms after cue onset and distinguished successful vs. unsuccessful recall. This increase in gamma power for successful recall was followed by a decrease in hippocampal alpha power (8 to 12 Hz). Intriguingly, the hippocampal gamma power increase marked the moment at which extrahippocampal activation patterns shifted from perceptual cue toward mnemonic target representations. In parallel, source-localized EEG alpha power revealed that the recall signal progresses from hippocampus to posterior parietal cortex and then to medial prefrontal cortex. Together, these results identify the hippocampus as the switchboard between perception and memory and elucidate the ensuing hippocampal–cortical dynamics supporting the recall process.

Pervasive cytoquakes in the actomyosin cortex across cell types and substrate stiffness

The actomyosin cytoskeleton enables cells to resist deformation, crawl, change their shape and sense their surroundings. Despite decades of study, how its molecular constituents can assemble together to form a network with the observed mechanics of cells remains poorly understood. Recently, it has been shown that the actomyosin cortex of quiescent cells can undergo frequent, abrupt reconfigurations and displacements, called cytoquakes. Notably, such fluctuations are not predicted by current physical models of actomyosin networks, and their prevalence across cell types and mechanical environments has not previously been studied. Using micropost array detectors, we have performed high-resolution measurements of the dynamic mechanical fluctuations of cells’ actomyosin cortex and stress fiber networks. This reveals cortical dynamics dominated by cytoquakes—intermittent events with a fat-tailed distribution of displacements, sometimes spanning microposts separated by 4 μm, in all cell types studied. These included 3T3 fibroblasts, where cytoquakes persisted over substrate stiffnesses spanning the tissue-relevant range of 4.3 kPa–17 kPa, and primary neonatal rat cardiac fibroblasts and myofibroblasts, human embryonic kidney cells and human bone osteosarcoma epithelial (U2OS) cells, where cytoquakes were observed on substrates in the same stiffness range. Overall, these findings suggest that the cortex self-organizes into a marginally stable mechanical state whose physics may contribute to cell mechanical properties, active behavior and mechanosensing.

Polymer skulls with integrated transparent electrode arrays for cortex-wide opto-electrophysiological recordings

Electrophysiology and optical imaging provide complementary neural sensing capabilities; electrophysiological recordings have the highest temporal resolution, while optical imaging allows recording the activities of genetically defined populations at high spatial resolution. Combining these complementary, yet orthogonal modalities to perform simultaneous large-scale, multimodal sensing of neural activity across multiple brain regions would be very powerful. Here we show that transparent, inkjet-printed electrocorticography (ECoG) electrode arrays can be seamlessly integrated with morphologically conformant transparent polymer skulls for multimodal recordings across the cortex. These eSee-Shells, were implanted on transgenic mice expressing the Ca2+ indicator GCaMP6f in cortical excitatory cells and provided a robust opto-electrophysiological interface for over 100 days. eSee-Shells enable simultaneous mesoscale Ca2+ imaging and ECoG acquisition under anesthesia as well as in awake animals presented with sensory stimuli. eSee-Shells further show sufficient clarity and transparency to observe single-cell Ca2+ signals directly below the electrodes and interconnects. Simultaneous multimodal measurement of cortical dynamics reveals changes in both ECoG and Ca2+ signals that depend on the behavioral state.

Traveling Waves in Quasi-One-Dimensional Neuronal Minicolumns

Traveling waves of neuronal activity in the cortex have been observed in vivo. These traveling waves have been correlated to various features of observed cortical dynamics, including spike timing variability and correlated fluctuations in neuron membrane potential. Although traveling waves are typically studied as either strictly one-dimensional or two-dimensional excitations, here we investigate the conditions for the existence of quasi-one-dimensional traveling waves that could be sustainable in parts of the brain containing cortical minicolumns. For that, we explore a quasi-one-dimensional network of heterogeneous neurons with a biologically influenced computational model of neuron dynamics and connectivity. We find that background stimulus reliably evokes traveling waves in networks with local connectivity between neurons. We also observe traveling waves in fully connected networks when a model for action potential propagation speed is incorporated. The biological properties of the neurons influence the generation and propagation of the traveling waves. Our quasi-one-dimensional model is not only useful for studying the basic properties of traveling waves in neuronal networks; it also provides a simplified representation of possible wave propagation in columnar or minicolumnar networks found in the cortex.

Deep neural networks and visuo-semantic models explain complementary components of human ventral-stream representational dynamics

Deep neural networks (DNNs) are promising models of the cortical computations supporting human object recognition. However, despite their ability to explain a significant portion of variance in neural data, the agreement between models and brain representational dynamics is far from perfect. Here, we address this issue by asking which representational features are currently unaccounted for in neural timeseries data, estimated for multiple areas of the human ventral stream via source-reconstructed magnetoencephalography (MEG) data. In particular, we focus on the ability of visuo-semantic models, consisting of human-generated labels of higher-level object features and categories, to explain variance beyond the explanatory power of DNNs alone. We report a gradual transition in the importance of visuo-semantic features from early to higher-level areas along the ventral stream. While early visual areas are better explained by DNN features, higher-level cortical dynamics are best accounted for by visuo-semantic models. These results suggest that current DNNs fail to fully capture the visuo-semantic features represented in higher-level human visual cortex and suggest a path towards more accurate models of ventral stream computations.