Saccade rate is associated with recall of items in working memory.

Working memory has been shown to rely on theta oscillations' phase synchronicity for item encoding and recall. At the same time, saccadic eye movements during visual exploration have been observed to trigger theta-phase resets, raising the question of whether the neuronal substrates of mnemonic processing rely on motor-evoked responses. To quantify the relationship between saccades and working memory load, we recorded eye tracking and behavioral data from human participants simultaneously performing an n-back Sternberg auditory task and a hue-based catch detection task. In addition to task-specific interference in performance, we also found that saccade rate was modulated by working memory load in the Sternberg task's preresponse stage. Our results support the possibility of interplay between saccades and hippocampal theta during working memory retrieval of items.

Theta-band phase locking during encoding leads to coordinated entorhinal-hippocampal replay.

Precisely timed interactions between hippocampal and cortical neurons during replay epochs are thought to support learning. Indeed, research has shown that replay is associated with heightened hippocampal-cortical synchrony. Yet many caveats remain in our understanding. Namely, it remains unclear how this offline synchrony comes about, whether it is specific to particular behavioral states, and how-if at all-it relates to learning. In this study, we sought to address these questions by analyzing coordination between CA1 cells and neurons of the deep layers of the medial entorhinal cortex (dMEC) while rats learned a novel spatial task. During movement, we found a subset of dMEC cells that were particularly locked to hippocampal LFP theta-band oscillations and that were preferentially coordinated with hippocampal replay during offline periods. Further, dMEC synchrony with CA1 replay peaked ∼10 ms after replay initiation in CA1, suggesting that the distributed replay reflects extra-hippocampal information propagation and is specific to "offline" periods. Finally, theta-modulated dMEC cells showed a striking experience-dependent increase in synchronization with hippocampal replay trajectories, mirroring the animals' acquisition of the novel task and coupling to the hippocampal local field. Together, these findings provide strong support for the hypothesis that synergistic hippocampal-cortical replay supports learning and highlights phase locking to hippocampal theta oscillations as a potential mechanism by which such cross-structural synchrony comes about.

Epistemic Autonomy: Self-supervised Learning in the Mammalian Hippocampus.

Biological cognition is based on the ability to autonomously acquire knowledge, or epistemic autonomy. Such self-supervision is largely absent in artificial neural networks (ANN) because they depend on externally set learning criteria. Yet training ANN using error backpropagation has created the current revolution in artificial intelligence, raising the question of whether the epistemic autonomy displayed in biological cognition can be achieved with error backpropagation-based learning. We present evidence suggesting that the entorhinal-hippocampal complex combines epistemic autonomy with error backpropagation. Specifically, we propose that the hippocampus minimizes the error between its input and output signals through a modulatory counter-current inhibitory network. We further discuss the computational emulation of this principle and analyze it in the context of autonomous cognitive systems.

Erratum: Entorhinal mismatch: A model of self-supervised learning in the hippocampus.

[This corrects the article DOI: 10.1016/j.isci.2021.102364.].

Entorhinal mismatch: A model of self-supervised learning in the hippocampus.

The hippocampal formation displays a wide range of physiological responses to different spatial manipulations of the environment. However, very few attempts have been made to identify core computational principles underlying those hippocampal responses. Here, we capitalize on the observation that the entorhinal-hippocampal complex (EHC) forms a closed loop and projects inhibitory signals "countercurrent" to the trisynaptic pathway to build a self-supervised model that learns to reconstruct its own inputs by error backpropagation. The EHC is then abstracted as an autoencoder, with the hidden layers acting as an information bottleneck. With the inputs mimicking the firing activity of lateral and medial entorhinal cells, our model is shown to generate place cells and to respond to environmental manipulations as observed in rodent experiments. Altogether, we propose that the hippocampus builds conjunctive compressed representations of the environment by learning to reconstruct its own entorhinal inputs via gradient descent.

Modulating grid cell scale and intrinsic frequencies via slow high-threshold conductances: A simplified model.

Grid cells in the medial entorhinal cortex (MEC) have known spatial periodic firing fields which provide a metric for the representation of self-location and path planning. The hexagonal tessellation pattern of grid cells scales up progressively along the MEC's layer II dorsal-to-ventral axis. This scaling gradient has been hypothesized to originate either from inter-population synaptic dynamics as postulated by attractor networks, or from projected theta frequency waves to different axis levels, as in oscillatory models. Alternatively, cellular dynamics and specifically slow high-threshold conductances have been proposed to have an impact on the grid cell scale. To test the hypothesis that intrinsic hyperpolarization-activated cation currents account for both the scaled gradient and the oscillatory frequencies observed along the dorsal-to-ventral axis, we have modeled and analyzed data from a population of grid cells simulated with spiking neurons interacting through low-dimensional attractor dynamics. We observed that the intrinsic neuronal membrane properties of simulated cells were sufficient to induce an increase in grid scale and potentiate differences in the membrane potential oscillatory frequency. Overall, our results suggest that the after-spike dynamics of cation currents may play a major role in determining the grid cells' scale and that oscillatory frequencies are a consequence of intrinsic cellular properties that are specific to different levels of the dorsal-to-ventral axis in the MEC layer II.

Human Vicarious Trial and Error Is Predictive of Spatial Navigation Performance.

When learning new environments, rats often pause at decision points and look back and forth over their possible trajectories as if they were imagining the future outcome of their actions, a behavior termed "Vicarious trial and error" (VTE). As the animal learns the environmental configuration, rats change from deliberative to habitual behavior, and VTE tends to disappear, suggesting a functional relevance in the early stages of learning. Despite the extensive research on spatial navigation, learning and VTE in the rat model, fewer studies have focused on humans. Here, we tested whether head-scanning behaviors that humans typically exhibit during spatial navigation are as predictive of spatial learning as in the rat. Subjects performed a goal-oriented virtual navigation task in a symmetric environment. Spatial learning was assessed through the analysis of trajectories, timings, and head orientations, under habitual and deliberative spatial navigation conditions. As expected, we found that trajectory length and duration decreased with the trial number, implying that subjects learned the spatial configuration of the environment over trials. Interestingly, IdPhi (a standard metric of VTE) also decreased with the trial number, suggesting that humans benefit from the same head-orientation scanning behavior as rats at spatial decision-points. Moreover, IdPhi captured exclusively at the first decision-point of each trial, was correlated with trial trajectory duration and length. Our findings demonstrate that in VTE is a signature of the stage of spatial learning in humans, and can be used to predict performance in navigation tasks with high accuracy.

Size Matters: How Scaling Affects the Interaction between Grid and Border Cells.

Many hippocampal cell types are characterized by a progressive increase in scale along the dorsal-to-ventral axis, such as in the cases of head-direction, grid and place cells. Also located in the medial entorhinal cortex (MEC), border cells would be expected to benefit from such scale modulations. However, this phenomenon has not been experimentally observed. Grid cells in the MEC of mammals integrate velocity related signals to map the environment with characteristic hexagonal tessellation patterns. Due to the noisy nature of these input signals, path integration processes tend to accumulate errors as animals explore the environment, leading to a loss of grid-like activity. It has been suggested that border-to-grid cells' associations minimize the accumulated grid cells' error when rodents explore enclosures. Thus, the border-grid interaction for error minimization is a suitable scenario to study the effects of border cell scaling within the context of spatial representation. In this study, we computationally address the question of (i) border cells' scale from the perspective of their role in maintaining the regularity of grid cells' firing fields, as well as (ii) what are the underlying mechanisms of grid-border associations relative to the scales of both grid and border cells. Our results suggest that for optimal contribution to grid cells' error minimization, border cells should express smaller firing fields relative to those of the associated grid cells, which is consistent with the hypothesis of border cells functioning as spatial anchoring signals.

An embodied biologically constrained model of foraging: from classical and operant conditioning to adaptive real-world behavior in DAC-X.

Animals successfully forage within new environments by learning, simulating and adapting to their surroundings. The functions behind such goal-oriented behavior can be decomposed into 5 top-level objectives: 'how', 'why', 'what', 'where', 'when' (H4W). The paradigms of classical and operant conditioning describe some of the behavioral aspects found in foraging. However, it remains unclear how the organization of their underlying neural principles account for these complex behaviors. We address this problem from the perspective of the Distributed Adaptive Control theory of mind and brain (DAC) that interprets these two paradigms as expressing properties of core functional subsystems of a layered architecture. In particular, we propose DAC-X, a novel cognitive architecture that unifies the theoretical principles of DAC with biologically constrained computational models of several areas of the mammalian brain. DAC-X supports complex foraging strategies through the progressive acquisition, retention and expression of task-dependent information and associated shaping of action, from exploration to goal-oriented deliberation. We benchmark DAC-X using a robot-based hoarding task including the main perceptual and cognitive aspects of animal foraging. We show that efficient goal-oriented behavior results from the interaction of parallel learning mechanisms accounting for motor adaptation, spatial encoding and decision-making. Together, our results suggest that the H4W problem can be solved by DAC-X building on the insights from the study of classical and operant conditioning. Finally, we discuss the advantages and limitations of the proposed biologically constrained and embodied approach towards the study of cognition and the relation of DAC-X to other cognitive architectures.