Canonical Microcircuit Research Articles

Reliability of dynamic causal modelling of resting-state magnetoencephalography.

This study assesses the reliability of resting-state dynamic causal modelling (DCM) of magnetoencephalography (MEG) under conductance-based canonical microcircuit models, in terms of both posterior parameter estimates and model evidence. We use resting-state MEG data from two sessions, acquired 2 weeks apart, from a cohort with high between-subject variance arising from Alzheimer's disease. Our focus is not on the effect of disease, but on the reliability of the methods (as within-subject between-session agreement), which is crucial for future studies of disease progression and drug intervention. To assess the reliability of first-level DCMs, we compare model evidence associated with the covariance among subject-specific free energies (i.e., the 'quality' of the models) with versus without interclass correlations. We then used parametric empirical Bayes (PEB) to investigate the differences between the inferred DCM parameter probability distributions at the between subject level. Specifically, we examined the evidence for or against parameter differences (i) within-subject, within-session, and between-epochs; (ii) within-subject between-session; and (iii) within-site between-subjects, accommodating the conditional dependency among parameter estimates. We show that for data acquired close in time, and under similar circumstances, more than 95% of inferred DCM parameters are unlikely to differ, speaking to mutual predictability over sessions. Using PEB, we show a reciprocal relationship between a conventional definition of 'reliability' and the conditional dependency among inferred model parameters. Our analyses confirm the reliability and reproducibility of the conductance-based DCMs for resting-state neurophysiological data. In this respect, the implicit generative modelling is suitable for interventional and longitudinal studies of neurological and psychiatric disorders.

Exploring biological challenges in building a thinking machine

This article presents a transdisciplinary analysis of the challenges in fusing neuroscience concepts with artificial intelligence (AI) to create AI systems inspired by biological cognition. We explore the structural and functional disparities between the neocortex’s canonical microcircuits and existing AI models, focusing on architectural differences, learning mechanisms, and energy efficiency. The discussion extends to adapting non-goal-oriented learning and dynamic neuronal connections from biological brains to enhance AI’s flexibility and efficiency. This work underscores the potential of neuroscientific insights to revolutionize AI development, advocating for a paradigm shift towards more adaptable and brain-like AI systems. We conclude that there is major room for bioinspiration by focusing on developing architecture, objective functions, and learning rules using a local instead of a global approach.

Exploring the Architectural Biases of the Canonical Cortical Microcircuit.

The cortex plays a crucial role in various perceptual and cognitive functions, driven by its basic unit, the canonical cortical microcircuit. Yet, we remain short of a framework that definitively explains the structure-function relationships of this fundamental neuroanatomical motif. To better understand how physical substrates of cortical circuitry facilitate their neuronal dynamics, we employ a computational approach using recurrent neural networks and representational analyses. We examine the differences manifested by the inclusion and exclusion of biologically-motivated inter-areal laminar connections on the computational roles of different neuronal populations in the microcircuit of two hierarchically-related areas, throughout learning. Our findings show that the presence of feedback connections correlates with the functional modularization of cortical populations in different layers, and provides the microcircuit with a natural inductive bias to differentiate expected and unexpected inputs at initialization. Furthermore, when testing the effects of training the microcircuit and its variants with a predictive-coding inspired strategy, we find that doing so helps better encode noisy stimuli in areas of the cortex that receive feedback, all of which combine to suggest evidence for a predictive-coding mechanism serving as an intrinsic operative logic in the cortex.

Top–down modulation in canonical cortical circuits with short-term plasticity

Cortical dynamics and computations are strongly influenced by diverse GABAergic interneurons, including those expressing parvalbumin (PV), somatostatin (SST), and vasoactive intestinal peptide (VIP). Together with excitatory (E) neurons, they form a canonical microcircuit and exhibit counterintuitive nonlinear phenomena. One instance of such phenomena is response reversal, whereby SST neurons show opposite responses to top-down modulation via VIP depending on the presence of bottom-up sensory input, indicating that the network may function in different regimes under different stimulation conditions. Combining analytical and computational approaches, we demonstrate that model networks with multiple interneuron subtypes and experimentally identified short-term plasticity mechanisms can implement response reversal. Surprisingly, despite not directly affecting SST and VIP activity, PV-to-E short-term depression has a decisive impact on SST response reversal. We show how response reversal relates to inhibition stabilization and the paradoxical effect in the presence of several short-term plasticity mechanisms demonstrating that response reversal coincides with a change in the indispensability of SST for network stabilization. In summary, our work suggests a role of short-term plasticity mechanisms in generating nonlinear phenomena in networks with multiple interneuron subtypes and makes several experimentally testable predictions.

Agranular frontal cortical microcircuit underlying cognitive control in macaques.

The error-related negativity and an N2-component recorded over medial frontal cortex index core functions of cognitive control. While they are known to originate from agranular frontal areas, the underlying microcircuit mechanisms remain elusive. Most insights about microcircuit function have been derived from variations of the so-called canonical microcircuit model. These microcircuit architectures are based extensively on studies from granular sensory cortical areas in monkeys, cats, and rodents. However, evidence has shown striking cytoarchitectonic differences across species and differences in the functional relationships across cortical layers in agranular compared to granular sensory areas. In this minireview, we outline a tentative microcircuit model underlying cognitive control in the agranular frontal cortex of primates. The model incorporates the main GABAergic interneuron subclasses with specific laminar arrangements and target regions on pyramidal cells. We emphasize the role of layer 5 pyramidal cells in error and conflict detection. We offer several specific questions necessary for creating a specific intrinsic microcircuit model of the agranular frontal cortex.

Neural heterogeneity controls computations in spiking neural networks.

The brain is composed of complex networks of interacting neurons that express considerable heterogeneity in their physiology and spiking characteristics. How does this neural heterogeneity influence macroscopic neural dynamics, and how might it contribute to neural computation? In this work, we use a mean-field model to investigate computation in heterogeneous neural networks, by studying how the heterogeneity of cell spiking thresholds affects three key computational functions of a neural population: the gating, encoding, and decoding of neural signals. Our results suggest that heterogeneity serves different computational functions in different cell types. In inhibitory interneurons, varying the degree of spike threshold heterogeneity allows them to gate the propagation of neural signals in a reciprocally coupled excitatory population. Whereas homogeneous interneurons impose synchronized dynamics that narrow the dynamic repertoire of the excitatory neurons, heterogeneous interneurons act as an inhibitory offset while preserving excitatory neuron function. Spike threshold heterogeneity also controls the entrainment properties of neural networks to periodic input, thus affecting the temporal gating of synaptic inputs. Among excitatory neurons, heterogeneity increases the dimensionality of neural dynamics, improving the network's capacity to perform decoding tasks. Conversely, homogeneous networks suffer in their capacity for function generation, but excel at encoding signals via multistable dynamic regimes. Drawing from these findings, we propose intra-cell-type heterogeneity as a mechanism for sculpting the computational properties of local circuits of excitatory and inhibitory spiking neurons, permitting the same canonical microcircuit to be tuned for diverse computational tasks.

Microcircuit failure in STXBP1 encephalopathy leads to hyperexcitability

De novo mutations in STXBP1 are among the most prevalent causes of neurodevelopmental disorders and lead to haploinsufficiency, cortical hyperexcitability, epilepsy, and other symptoms in people with mutations. Given that Munc18-1, the protein encoded by STXBP1, is essential for excitatory and inhibitory synaptic transmission, it is currently not understood why mutations cause hyperexcitability. We find that overall inhibition in canonical feedforward microcircuits is defective in a P15-22 mouse model for Stxbp1 haploinsufficiency. Unexpectedly, we find that inhibitory synapses formed by parvalbumin-positive interneurons were largely unaffected. Instead, excitatory synapses fail to recruit inhibitory interneurons. Modeling confirms that defects in the recruitment of inhibitory neurons cause hyperexcitation. CX516, an ampakine that enhances excitatory synapses, restores interneuron recruitment and prevents hyperexcitability. These findings establish deficits in excitatory synapses in microcircuits as a key underlying mechanism for cortical hyperexcitability in a mouse model of Stxbp1 disorder and identify compounds enhancing excitation as a direction for therapy.

Neuropharmacological computational analysis of longitudinal electroencephalograms in clozapine-treated patients with schizophrenia using hierarchical dynamic causal modeling

The hierarchical characteristics of the brain are prominent in the pharmacological treatment of psychiatric diseases, primarily targeting cellular receptors that extend upward to intrinsic connectivity within a region, interregional connectivity, and, consequently, clinical observations such as an electroencephalogram (EEG). To understand the long-term effects of neuropharmacological intervention on neurobiological properties at different hierarchical levels, we explored long-term changes in neurobiological parameters of an N-methyl-D-aspartate canonical microcircuit model (CMM-NMDA) in the default mode network (DMN) and auditory hallucination network (AHN) using dynamic causal modeling of longitudinal EEG in clozapine-treated patients with schizophrenia. The neurobiological properties of the CMM-NMDA model associated with symptom improvement in schizophrenia were found across hierarchical levels, from a reduced membrane capacity of the deep pyramidal cell and intrinsic connectivity with the inhibitory population in DMN and intrinsic and extrinsic connectivity in AHN. The medication duration mainly affects the intrinsic connectivity and NMDA time constant in DMN. Virtual perturbation analysis specified the contribution of each parameter to the cross-spectral density (CSD) of the EEG, particularly intrinsic connectivity and membrane capacitances for CSD frequency shifts and progression. It further reveals that excitatory and inhibitory connectivity complements frequency-specific CSD changes, notably the alpha frequency band in DMN. Positive and negative synergistic interactions exist between neurobiological properties primarily within the same region in patients treated with clozapine. The current study shows how computational neuropharmacology helps explore the multiscale link between neurobiological properties and clinical observations and understand the long-term mechanism of neuropharmacological intervention reflected in clinical EEG.

Global dynamics of neural mass models

Neural mass models are used to simulate cortical dynamics and to explain the electrical and magnetic fields measured using electro- and magnetoencephalography. Simulations evince a complex phase-space structure for these kinds of models; including stationary points and limit cycles and the possibility for bifurcations and transitions among different modes of activity. This complexity allows neural mass models to describe the itinerant features of brain dynamics. However, expressive, nonlinear neural mass models are often difficult to fit to empirical data without additional simplifying assumptions: e.g., that the system can be modelled as linear perturbations around a fixed point. In this study we offer a mathematical analysis of neural mass models, specifically the canonical microcircuit model, providing analytical solutions describing slow changes in the type of cortical activity, i.e. dynamical itinerancy. We derive a perturbation analysis up to second order of the phase flow, together with adiabatic approximations. This allows us to describe amplitude modulations in a relatively simple mathematical format providing analytic proof-of-principle for the existence of semi-stable states of cortical dynamics at the scale of a cortical column. This work allows for model inversion of neural mass models, not only around fixed points, but over regions of phase space that encompass transitions among semi or multi-stable states of oscillatory activity. Crucially, these theoretical results speak to model inversion in the context of multiple semi-stable brain states, such as the transition between interictal, pre-ictal and ictal activity in epilepsy.

Dynamic causal modelling shows a prominent role of local inhibition in alpha power modulation in higher visual cortex.

During resting-state EEG recordings, alpha activity is more prominent over the posterior cortex in eyes-closed (EC) conditions compared to eyes-open (EO). In this study, we characterized the difference in spectra between EO and EC conditions using dynamic causal modelling. Specifically, we investigated the role of intrinsic and extrinsic connectivity-within the visual cortex-in generating EC-EO alpha power differences over posterior electrodes. The primary visual cortex (V1) and the bilateral middle temporal visual areas (V5) were equipped with bidirectional extrinsic connections using a canonical microcircuit. The states of four intrinsically coupled subpopulations-within each occipital source-were also modelled. Using Bayesian model selection, we tested whether modulations of the intrinsic connections in V1, V5 or extrinsic connections (or a combination thereof) provided the best evidence for the data. In addition, using parametric empirical Bayes (PEB), we estimated group averages under the winning model. Bayesian model selection showed that the winning model contained both extrinsic connectivity modulations, as well as intrinsic connectivity modulations in all sources. The PEB analysis revealed increased extrinsic connectivity during EC. Overall, we found a reduction in the inhibitory intrinsic connections during EC. The results suggest that the intrinsic modulations in V5 played the most important role in producing EC-EO alpha differences, suggesting an intrinsic disinhibition in higher order visual cortex, during EC resting state.

Neurophysiological consequences of synapse loss in progressive supranuclear palsy.

Synaptic loss occurs early in many neurodegenerative diseases and contributes to cognitive impairment even in the absence of gross atrophy. Currently, for human disease there are few formal models to explain how cortical networks underlying cognition are affected by synaptic loss. We advocate that biophysical models of neurophysiology offer both a bridge from preclinical to clinical models of pathology and quantitative assays for experimental medicine. Such biophysical models can also disclose hidden neuronal dynamics generating neurophysiological observations such as EEG and magnetoencephalography. Here, we augment a biophysically informed mesoscale model of human cortical function by inclusion of synaptic density estimates as captured by 11C-UCB-J PET, and provide insights into how regional synapse loss affects neurophysiology. We use the primary tauopathy of progressive supranuclear palsy (Richardson's syndrome) as an exemplar condition, with high clinicopathological correlations. Progressive supranuclear palsy causes a marked change in cortical neurophysiology in the presence of mild cortical atrophy and is associated with a decline in cognitive functions associated with the frontal lobe. Using parametric empirical Bayesian inversion of a conductance-based canonical microcircuit model of magnetoencephalography data, we show that the inclusion of regional synaptic density-as a subject-specific prior on laminar-specific neuronal populations-markedly increases model evidence. Specifically, model comparison suggests that a reduction in synaptic density in inferior frontal cortex affects superficial and granular layer glutamatergic excitation. This predicted individual differences in behaviour, demonstrating the link between synaptic loss, neurophysiology and cognitive deficits. The method we demonstrate is not restricted to progressive supranuclear palsy or the effects of synaptic loss: such pathology-enriched dynamic causal models can be used to assess the mechanisms of other neurological disorders, with diverse non-invasive measures of pathology, and is suitable to test the effects of experimental pharmacology.

Laminar-specific functional connectivity mapping with multi-slice line-scanning fMRI.

Despite extensive studies detecting laminar functional magnetic resonance imaging (fMRI) signals to illustrate the canonical microcircuit, the spatiotemporal characteristics of laminar-specific information flow across cortical regions remain to be fully investigated in both evoked and resting conditions at different brain states. Here, we developed a multislice line-scanning fMRI (MS-LS) method to detect laminar fMRI signals in adjacent cortical regions with high spatial (50 μm) and temporal resolution (100 ms) in anesthetized rats. Across different trials, we detected either laminar-specific positive or negative blood-oxygen-level-dependent (BOLD) responses in the surrounding cortical region adjacent to the most activated cortex under the evoked condition. Specifically, in contrast to typical Layer (L) 4 correlation across different regions due to the thalamocortical projections for trials with positive BOLD, a strong correlation pattern specific in L2/3 was detected for trials with negative BOLD in adjacent regions, which indicated brain state-dependent laminar-fMRI responses based on corticocortical interaction. Also, in resting-state (rs-) fMRI study, robust lag time differences in L2/3, 4, and 5 across multiple cortices represented the low-frequency rs-fMRI signal propagation from caudal to rostral slices. In summary, our study provided a unique laminar fMRI mapping scheme to better characterize trial-specific intra- and inter-laminar functional connectivity in evoked and resting-state MS-LS.

Canonical microcircuits and the compressibility of neural connectomes

A prominent hypothesis for how the brain's wiring is organized is that it is made up of stereotyped “canonical” microcircuits. Relying on a restricted set of repeated circuits, optimized for specific functions, may provide biologically advantageous inductive biases for efficient learning and help encode innate behaviors. Information theory furthermore tells us that the presence of such statistically regular circuits, termed motifs, would make the brain's wiring diagram, the connectome, compressible.

Synaptic connectivity to L2/3 of primary visual cortex measured by two-photon optogenetic stimulation.

Understanding cortical microcircuits requires thorough measurement of physiological properties of synaptic connections formed within and between diverse subclasses of neurons. Towards this goal, we combined spatially precise optogenetic stimulation with multicellular recording to deeply characterize intralaminar and translaminar monosynaptic connections to supragranular (L2/3) neurons in the mouse visual cortex. The reliability and specificity of multiphoton optogenetic stimulation were measured across multiple Cre lines, and measurements of connectivity were verified by comparison to paired recordings and targeted patching of optically identified presynaptic cells. With a focus on translaminar pathways, excitatory and inhibitory synaptic connections from genetically defined presynaptic populations were characterized by their relative abundance, spatial profiles, strength, and short-term dynamics. Consistent with the canonical cortical microcircuit, layer 4 excitatory neurons and interneurons within L2/3 represented the most common sources of input to L2/3 pyramidal cells. More surprisingly, we also observed strong excitatory connections from layer 5 intratelencephalic neurons and potent translaminar inhibition from multiple interneuron subclasses. The hybrid approach revealed convergence to and divergence from excitatory and inhibitory neurons within and across cortical layers. Divergent excitatory connections often spanned hundreds of microns of horizontal space. In contrast, divergent inhibitory connections were more frequently measured from postsynaptic targets near each other.

Information flow in the rat thalamo-cortical system: spontaneous vs. stimulus-evoked activities

The interaction between the thalamus and sensory cortex plays critical roles in sensory processing. Previous studies have revealed pathway-specific synaptic properties of thalamo-cortical connections. However, few studies to date have investigated how each pathway routes moment-to-moment information. Here, we simultaneously recorded neural activity in the auditory thalamus (or ventral division of the medial geniculate body; MGv) and primary auditory cortex (A1) with a laminar resolution in anesthetized rats. Transfer entropy (TE) was used as an information theoretic measure to operationalize “information flow”. Our analyses confirmed that communication between the thalamus and cortex was strengthened during presentation of auditory stimuli. In the resting state, thalamo-cortical communications almost disappeared, whereas intracortical communications were strengthened. The predominant source of information was the MGv at the onset of stimulus presentation and layer 5 during spontaneous activity. In turn, MGv was the major recipient of information from layer 6. TE suggested that a small but significant population of MGv-to-A1 pairs was “information-bearing,” whereas A1-to-MGv pairs typically exhibiting small effects played modulatory roles. These results highlight the capability of TE analyses to unlock novel avenues for bridging the gap between well-established anatomical knowledge of canonical microcircuits and physiological correlates via the concept of dynamic information flow.

Laminar processing of numerosity supports a canonical cortical microcircuit in human parietal cortex

As neural signals travel through the visual hierarchy, spatial precision decreases and specificity for stimulus features increases.1-4 A similar hierarchy has been found for laminar processing in V1, where information from the thalamus predominantly targets the central layers, while spatial precision decreases and feature specificity increases toward superficial and deeper layers.5-17 This laminar processing scheme is proposed to represent a canonical cortical microcircuit that is similar across the cortex.11,18-21 Here, we go beyond early visual cortex and investigate whether processing of numerosity (the set size of a group of items) across cortical depth in the parietal association cortex follows this hypothesis. Numerosity processing is implicated in many tasks such as multiple object tracking,22 mathematics,23-25 decision making,26 and dividing attention.27 Neurons in the parietal association cortex are tuned to numerosity, with both a preferred numerosity tuning and tuning width (i.e., specificity).28-30 We quantified preferred numerosity responses across cortical depth in the parietal association cortex with ultra-high field fMRI and population receptive field-based numerosity modeling.1,28,31 We find that numerosity responses sharpen, i.e., become increasingly specific, moving away from the central layers. This suggests that the laminar processing scheme for numerosity processing in the parietal cortex is similar to primary visual cortex, providing support for the canonical cortical microcircuit hypothesis beyond primary visual cortex.

Stimulus Feature-Specific Information Flow Along the Columnar Cortical Microcircuit Revealed by Multivariate Laminar Spiking Analysis.

Most of the mammalian neocortex is comprised of a highly similar anatomical structure, consisting of a granular cell layer between superficial and deep layers. Even so, different cortical areas process different information. Taken together, this suggests that cortex features a canonical functional microcircuit that supports region-specific information processing. For example, the primate primary visual cortex (V1) combines the two eyes' signals, extracts stimulus orientation, and integrates contextual information such as visual stimulation history. These processes co-occur during the same laminar stimulation sequence that is triggered by the onset of visual stimuli. Yet, we still know little regarding the laminar processing differences that are specific to each of these types of stimulus information. Univariate analysis techniques have provided great insight by examining one electrode at a time or by studying average responses across multiple electrodes. Here we focus on multivariate statistics to examine response patterns across electrodes instead. Specifically, we applied multivariate pattern analysis (MVPA) to linear multielectrode array recordings of laminar spiking responses to decode information regarding the eye-of-origin, stimulus orientation, and stimulus repetition. MVPA differs from conventional univariate approaches in that it examines patterns of neural activity across simultaneously recorded electrode sites. We were curious whether this added dimensionality could reveal neural processes on the population level that are challenging to detect when measuring brain activity without the context of neighboring recording sites. We found that eye-of-origin information was decodable for the entire duration of stimulus presentation, but diminished in the deepest layers of V1. Conversely, orientation information was transient and equally pronounced along all layers. More importantly, using time-resolved MVPA, we were able to evaluate laminar response properties beyond those yielded by univariate analyses. Specifically, we performed a time generalization analysis by training a classifier at one point of the neural response and testing its performance throughout the remaining period of stimulation. Using this technique, we demonstrate repeating (reverberating) patterns of neural activity that have not previously been observed using standard univariate approaches.

Layer 4 Gates Plasticity in Visual Cortex Independent of a Canonical Microcircuit

Disrupting binocular vision during a developmental critical period can yield enduring changes to ocular dominance (OD) in primary visual cortex (V1). Here we investigated how this experience-dependent plasticity is coordinated within the laminar circuitry of V1 by deleting separately in each cortical layer (L) a gene required to close the critical period, nogo-66 receptor (ngr1). Deleting ngr1 in excitatory neurons in L4, but not in L2/3, L5, or L6, prevented closure of the critical period, and adult mice remained sensitive to brief monocular deprivation. Intracortical disinhibition, but not thalamocortical disinhibition, accompanied this OD plasticity. Both juvenile wild-type mice and adult mice lacking ngr1 in L4 displayed OD plasticity that advanced more rapidly L4 than L2/3 or L5. Interestingly, blocking OD plasticity in L2/3 with the drug AM-251 did not impair OD plasticity in L5. We propose that L4 restricts disinhibition and gates OD plasticity independent of a canonical cortical microcircuit.

Dynamic Recovery from Depression Enables Rate Encoding in Inhibitory Synapses.

SummaryParvalbumin-expressing fast-spiking interneurons (PV-INs) control network firing and the gain of cortical response to sensory stimulation. Crucial for these functions, PV-INs can sustain high-frequency firing with no accommodation. However, PV-INs also exhibit short-term depression (STD) during sustained activation, largely due to the depletion of synaptic resources (vesicles). In most synapses the rate of replenishment of depleted vesicles is constant, determining an inverse relationship between depression levels and the activation rate, which theoretically, severely limits rate-coding capabilities. We examined STD of the PV-IN to pyramidal cell synapse in the mouse visual cortex and found that in these synapses the recovery from depression is not constant but increases linearly with the frequency of use. By combining modeling, dynamic clamp, and optogenetics, we demonstrated that this recovery enables PV-INs to reduce pyramidal cell firing in a linear manner, which theoretically is crucial for controlling the gain of cortical visual responses.

The Synaptic Organization of Layer 6 Circuits Reveals Inhibition as a Major Output of a Neocortical Sublamina

The canonical cortical microcircuit has principally been defined by interlaminar excitatory connections among the six layers of the neocortex. However, excitatory neurons in layer 6 (L6), a layer whose functional organization is poorly understood, form relatively rare synaptic connections with other cortical excitatory neurons. Here, we show that the vast majority of parvalbumin inhibitory neurons in a sublamina within L6 send axons through the cortical layers toward the pia. These interlaminar inhibitory neurons receive local synaptic inputs from both major types of L6 excitatory neurons and receive stronger input from thalamocortical afferents than do neighboring pyramidal neurons. The distribution of these interlaminar interneurons and their synaptic connectivity further support a functional subdivision within the standard six layers of the cortex. Positioned to integrate local and long-distance inputs in this sublayer, these interneurons generate an inhibitory interlaminar output. These findings call for a revision to the canonical cortical microcircuit.