Neurophysiologically-informed markers of individual variability and pharmacological manipulation of human cortical gamma

Abstract

The ability to quantify synaptic function at the level of cortical microcircuits from non-invasive data would be enormously useful in the study of neuronal processing in humans and the pathophysiology that attends many neuropsychiatric disorders. Here, we provide proof of principle that one can estimate inter-and intra-laminar interactions among specific neuronal populations using induced gamma responses in the visual cortex of human subjects – using dynamic causal modelling based upon the canonical microcircuit (CMC; a simplistic model of a cortical column). Using variability in induced (spectral) responses over a large cohort of normal subjects, we find that the predominant determinants of gamma responses rest on recurrent and intrinsic connections between superficial pyramidal cells and inhibitory interneurons. Furthermore, variations in beta responses were mediated by inter-subject differences in the intrinsic connections between deep pyramidal cells and inhibitory interneurons. Interestingly, we also show that increasing the self-inhibition of superficial pyramidal cells suppresses the amplitude of gamma activity, while increasing its peak frequency. This systematic and nonlinear relationship was only disclosed by modelling the causes of induced responses. Crucially, we were able to validate this form of neurophysiological phenotyping by showing a selective effect of the GABA re-uptake inhibitor tiagabine on the rate constants of inhibitory interneurons. Remarkably, we were able to recover the pharmacodynamics of this effect over the course of several hours on a per subject basis. These findings speak to the possibility of measuring population specific synaptic function – and its response to pharmacological intervention – to provide subject-specific biomarkers of mesoscopic neuronal processes using non-invasive data. Finally, our results demonstrate that, using the CMC as a proxy, the synaptic mechanisms that underlie the gain control of neuronal message passing within and between different levels of cortical hierarchies may now be amenable to quantitative study using non-invasive (MEG) procedures.

Peak visual gamma frequency is modified across the healthy menstrual cycle

“…Dynamic causal modeling for steady‐state responses (DCM‐SSR) was conducted using the methods of Shaw et al () which utilised a variation on the canonical microcircuit neural model (CMC). DCM uses a generative model comprising the neural‐mass model as well as an observation model, which when using EEG or MEG is typically a lead‐field weighting (Moran et al, ; Shaw et al, ). The neural mass model for the CMC comprises 4 types of interacting cell populations.…”

“…The neural mass model for the CMC comprises 4 types of interacting cell populations. For these the variation on the neural model employed by Shaw et al () includes six types of parameter, including time‐constants (T), local (G), and extrinsic (A) synaptic connectivity strengths, exogenous input (C) strength, delay (D), and presynaptic firing (S). Of particular interest is the modulation of intrinsic or local connectivity (parameter G) but also population time‐constants (parameter T) (Figure ), these and an overall gain parameter (L) are allowed to vary all other parameter types outlined above are fixed in the model.…”

“…Of particular interest in analyses of visually induced gamma oscillations is the ability to make inferences about intrinsic connectivity between populations of superficial pyramidal cells and inhibitory interneurons, potentially reflecting the neural dynamics underlying the generation of gamma oscillations (Buzsaki & Wang, ). Recently Shaw et al () developed a customized canonical microcircuit model that better reflects the synaptic physiology of V1 than the generic mass models used in previous DCM work, in order to study visually induced gamma oscillations which arise from V1. Using their model of V1, Shaw et al () were able to show significant modulation of parameter strength when visually induced gamma activity was modified by tiagabine.…”

“…Recently Shaw et al () developed a customized canonical microcircuit model that better reflects the synaptic physiology of V1 than the generic mass models used in previous DCM work, in order to study visually induced gamma oscillations which arise from V1. Using their model of V1, Shaw et al () were able to show significant modulation of parameter strength when visually induced gamma activity was modified by tiagabine. Furthermore, they were able to differentiate between laminar specific events including localizing the generation of gamma oscillations to primarily superficial layers through pyramidal‐interneuron loops.…”

“…The aim of this study was to quantify visual gamma oscillation parameters over the course of the menstrual cycle in healthy women. Furthermore, using the DCM approach developed by Shaw et al (), we sought to make inferences about functional changes in the GABA system that may contribute to any changes found. The study used EEG to record visually induced gamma oscillations in the early follicular phase compared to the mid luteal phase, when hormones are at their lowest compared to their highest levels.…”