Intrinsic neuronal dynamics predict distinct functional roles during working memory

Invariant timescale hierarchy across the cortical somatosensory network

The ability of cortical networks to integrate information from different sources is essential for cognitive processes. On one hand, sensory areas exhibit fast dynamics often phase-locked to stimulation; on the other hand, frontal lobe areas with slow response latencies to stimuli must integrate and maintain information for longer periods. Thus, cortical areas may require different timescales depending on their functional role. Studying the cortical somatosensory network while monkeys discriminated between two vibrotactile stimulus patterns, we found that a hierarchical order could be established across cortical areas based on their intrinsic timescales. Further, even though subareas (areas 3b, 1, and 2) of the primary somatosensory (S1) cortex exhibit analogous firing rate responses, a clear differentiation was observed in their timescales. Importantly, we observed that this inherent timescale hierarchy was invariant between task contexts (demanding vs. nondemanding). Even if task context severely affected neural coding in cortical areas downstream to S1, their timescales remained unaffected. Moreover, we found that these time constants were invariant across neurons with different latencies or coding. Although neurons had completely different dynamics, they all exhibited comparable timescales within each cortical area. Our results suggest that this measure is demonstrative of an inherent characteristic of each cortical area, is not a dynamical feature of individual neurons, and does not depend on task demands.

A Diversity of Intrinsic Timescales Underlie Neural Computations

Neural processing occurs across a range of temporal scales. To facilitate this, the brain uses fast-changing representations reflecting momentary sensory input alongside more temporally extended representations, which integrate across both short and long temporal windows. The temporal flexibility of these representations allows animals to behave adaptively. Short temporal windows facilitate adaptive responding in dynamic environments, while longer temporal windows promote the gradual integration of information across time. In the cognitive and motor domains, the brain sets overarching goals to be achieved within a long temporal window, which must be broken down into sequences of actions and precise movement control processed across much shorter temporal windows. Previous human neuroimaging studies and large-scale artificial network models have ascribed different processing timescales to different cortical regions, linking this to each region's position in an anatomical hierarchy determined by patterns of inter-regional connectivity. However, even within cortical regions, there is variability in responses when studied with single-neuron electrophysiology. Here, we review a series of recent electrophysiology experiments that demonstrate the heterogeneity of temporal receptive fields at the level of single neurons within a cortical region. This heterogeneity appears functionally relevant for the computations that neurons perform during decision-making and working memory. We consider anatomical and biophysical mechanisms that may give rise to a heterogeneity of timescales, including recurrent connectivity, cortical layer distribution, and neurotransmitter receptor expression. Finally, we reflect on the computational relevance of each brain region possessing a heterogeneity of neuronal timescales. We argue that this architecture is of particular importance for sensory, motor, and cognitive computations.

Reward signaling by the rodent medial frontal cortex

The anatomical relevance and functional significance of medial parts of the rodent frontal cortex have been intensely debated over the modern history of neuroscience. Early studies emphasized common functions among medial frontal regions in rodents and the dorsolateral prefrontal cortex of primates. Behavioral tasks emphasized memory-guided performance and persistent neural activity as a marker of working memory. Over time, it became clear that long-standing concerns about cross-species homology were justified and the view emerged that rodents are useful for understanding medial parts of the frontal cortex in primates, and not the dorsolateral prefrontal cortex. Here, we summarize a series of studies on the rodent medial frontal cortex that began with an interest in studying working memory in the perigenual prelimbic area and ended up studying reward processing in the medial orbital region. Our experiments revealed a role for a 4-8Hz "theta" rhythm in tracking engagement in the consumption of rewarding fluids and denoting the value of a given reward. Evidence for a functional differentiation between the rostral and caudal medial frontal cortex and its relationship to other frontal cortical areas is also discussed with the hope of motivating future work on this part of the cerebral cortex.

Strong inhibitory signaling underlies stable temporal dynamics and working memory in spiking neural networks

Cortical neurons process information on multiple timescales, and areas important for working memory (WM) contain neurons capable of integrating information over a long timescale. However, the underlying mechanisms for the emergence of neuronal timescales stable enough to support WM are unclear. By analyzing a spiking recurrent neural network model trained on a WM task and activity of single neurons in the primate prefrontal cortex, we show that the temporal properties of our model and the neural data are remarkably similar. Dissecting our recurrent neural network model revealed strong inhibitory-to-inhibitory connections underlying a disinhibitory microcircuit as a critical component for long neuronal timescales and WM maintenance. We also found that enhancing inhibitory-to-inhibitory connections led to more stable temporal dynamics and improved task performance. Finally, we show that a network with such microcircuitry can perform other tasks without disrupting its pre-existing timescale architecture, suggesting that strong inhibitory signaling underlies a flexible WM network.

Neuronal timescales are functionally dynamic and shaped by cortical microarchitecture

Complex cognitive functions such as working memory and decision-making require information maintenance over seconds to years, from transient sensory stimuli to long-term contextual cues. While theoretical accounts predict the emergence of a corresponding hierarchy of neuronal timescales, direct electrophysiological evidence across the human cortex is lacking. Here, we infer neuronal timescales from invasive intracranial recordings. Timescales increase along the principal sensorimotor-to-association axis across the entire human cortex, and scale with single-unit timescales within macaques. Cortex-wide transcriptomic analysis shows direct alignment between timescales and expression of excitation- and inhibition-related genes, as well as genes specific to voltage-gated transmembrane ion transporters. Finally, neuronal timescales are functionally dynamic: prefrontal cortex timescales expand during working memory maintenance and predict individual performance, while cortex-wide timescales compress with aging. Thus, neuronal timescales follow cytoarchitectonic gradients across the human cortex and are relevant for cognition in both short and long terms, bridging microcircuit physiology with macroscale dynamics and behavior.

Are intrinsic neural timescales related to sensory processing? Evidence from abnormal behavioral states

The brain exhibits a complex temporal structure which translates into a hierarchy of distinct neural timescales. An open question is how these intrinsic timescales are related to sensory or motor information processing and whether these dynamics have common patterns in different behavioral states. We address these questions by investigating the brain's intrinsic timescales in healthy controls, motor (amyotrophic lateral sclerosis, locked-in syndrome), sensory (anesthesia, unresponsive wakefulness syndrome), and progressive reduction of sensory processing (from awake states over N1, N2, N3). We employed a combination of measures from EEG resting-state data: auto-correlation window (ACW), power spectral density (PSD), and power-law exponent (PLE). Prolonged neural timescales accompanied by a shift towards slower frequencies were observed in the conditions with sensory deficits, but not in conditions with motor deficits. Our results establish that the spontaneous activity's intrinsic neural timescale is related to the neural capacity that specifically supports sensory rather than motor information processing in the healthy brain.

Predicting intelligence from brain gray matter volume

A positive association between brain size and intelligence is firmly established, but whether region-specific anatomical differences contribute to general intelligence remains an open question. Results from voxel-based morphometry (VBM) - one of the most widely used morphometric methods - have remained inconclusive so far. Here, we applied cross-validated machine learning-based predictive modeling to test whether out-of-sample prediction of individual intelligence scores is possible on the basis of voxel-wise gray matter volume. Features were derived from structural magnetic resonance imaging data (N = 308) using (a) a purely data-driven method (principal component analysis) and (b) a domain knowledge-based approach (atlas parcellation). When using relative gray matter (corrected for total brain size), only the atlas-based approach provided significant prediction, while absolute gray matter (uncorrected) allowed for above-chance prediction with both approaches. Importantly, in all significant predictions, the absolute error was relatively high, i.e., greater than ten IQ points, and in the atlas-based models, the predicted IQ scores varied closely around the sample mean. This renders the practical value even of statistically significant prediction results questionable. Analyses based on the gray matter of functional brain networks yielded significant predictions for the fronto-parietal network and the cerebellum. However, the mean absolute errors were not reduced in contrast to the global models, suggesting that general intelligence may be related more to global than region-specific differences in gray matter volume. More generally, our study highlights the importance of predictive statistical analysis approaches for clarifying the neurobiological bases of intelligence and provides important suggestions for future research using predictive modeling.

Neuronal spike-rate adaptation supports working memory in language processing

Language processing involves the ability to store and integrate pieces of information in working memory over short periods of time. According to the dominant view, information is maintained through sustained, elevated neural activity. Other work has argued that short-term synaptic facilitation can serve as a substrate of memory. Here we propose an account where memory is supported by intrinsic plasticity that downregulates neuronal firing rates. Single neuron responses are dependent on experience, and we show through simulations that these adaptive changes in excitability provide memory on timescales ranging from milliseconds to seconds. On this account, spiking activity writes information into coupled dynamic variables that control adaptation and move at slower timescales than the membrane potential. From these variables, information is continuously read back into the active membrane state for processing. This neuronal memory mechanism does not rely on persistent activity, excitatory feedback, or synaptic plasticity for storage. Instead, information is maintained in adaptive conductances that reduce firing rates and can be accessed directly without cued retrieval. Memory span is systematically related to both the time constant of adaptation and baseline levels of neuronal excitability. Interference effects within memory arise when adaptation is long lasting. We demonstrate that this mechanism is sensitive to context and serial order which makes it suitable for temporal integration in sequence processing within the language domain. We also show that it enables the binding of linguistic features over time within dynamic memory registers. This work provides a step toward a computational neurobiology of language.

Multiple timescales of neural dynamics and integration of task-relevant signals across cortex

A long-lasting challenge in neuroscience has been to find a set of principles that could be used to organize the brain into distinct areas with specific functions. Recent studies have proposed the orderly progression in the time constants of neural dynamics as an organizational principle of cortical computations. However, relationships between these timescales and their dependence on response properties of individual neurons are unknown, making it impossible to determine how mechanisms underlying such a computational principle are related to other aspects of neural processing. Here, we developed a comprehensive method to simultaneously estimate multiple timescales in neuronal dynamics and integration of task-relevant signals along with selectivity to those signals. By applying our method to neural and behavioral data during a dynamic decision-making task, we found that most neurons exhibited multiple timescales in their response, which consistently increased from parietal to prefrontal and cingulate cortex. While predicting rates of behavioral adjustments, these timescales were not correlated across individual neurons in any cortical area, resulting in independent parallel hierarchies of timescales. Additionally, none of these timescales depended on selectivity to task-relevant signals. Our results not only suggest the existence of multiple canonical mechanisms for increasing timescales of neural dynamics across cortex but also point to additional mechanisms that allow decorrelation of these timescales to enable more flexibility.

Hierarchical dynamics as a macroscopic organizing principle of the human brain

Multimodal evidence suggests that brain regions accumulate information over timescales that vary according to anatomical hierarchy. Thus, these experimentally defined "temporal receptive windows" are longest in cortical regions that are distant from sensory input. Interestingly, spontaneous activity in these regions also plays out over relatively slow timescales (i.e., exhibits slower temporal autocorrelation decay). These findings raise the possibility that hierarchical timescales represent an intrinsic organizing principle of brain function. Here, using resting-state functional MRI, we show that the timescale of ongoing dynamics follows hierarchical spatial gradients throughout human cerebral cortex. These intrinsic timescale gradients give rise to systematic frequency differences among large-scale cortical networks and predict individual-specific features of functional connectivity. Whole-brain coverage permitted us to further investigate the large-scale organization of subcortical dynamics. We show that cortical timescale gradients are topographically mirrored in striatum, thalamus, and cerebellum. Finally, timescales in the hippocampus followed a posterior-to-anterior gradient, corresponding to the longitudinal axis of increasing representational scale. Thus, hierarchical dynamics emerge as a global organizing principle of mammalian brains.

Stable and dynamic representations of value in the prefrontal cortex

Optimal decision-making requires that stimulus-value associations are kept up to date by constantly comparing the expected value of a stimulus with its experienced outcome. To do this, value information must be held in mind when a stimulus and outcome are separated in time. However, little is known about the neural mechanisms of working memory (WM) for value. Contradicting theories have suggested WM requires either persistent or transient neuronal activity, with stable or dynamic representations, respectively. To test these hypotheses, we recorded neuronal activity in the orbitofrontal and anterior cingulate cortex of two monkeys performing a valuation task. We found that features of all hypotheses were simultaneously present in prefrontal activity, and no single hypothesis was exclusively supported. Instead, mixed dynamics supported robust, time invariant value representations while also encoding the information in a temporally specific manner. We suggest that this hybrid coding is a critical mechanism supporting flexible cognitive abilities.

Multiscale Computation and Dynamic Attention in Biological and Artificial Intelligence

Biological and artificial intelligence (AI) are often defined by their capacity to achieve a hierarchy of short-term and long-term goals that require incorporating information over time and space at both local and global scales. More advanced forms of this capacity involve the adaptive modulation of integration across scales, which resolve computational inefficiency and explore-exploit dilemmas at the same time. Research in neuroscience and AI have both made progress towards understanding architectures that achieve this. Insight into biological computations come from phenomena such as decision inertia, habit formation, information search, risky choices and foraging. Across these domains, the brain is equipped with mechanisms (such as the dorsal anterior cingulate and dorsolateral prefrontal cortex) that can represent and modulate across scales, both with top-down control processes and by local to global consolidation as information progresses from sensory to prefrontal areas. Paralleling these biological architectures, progress in AI is marked by innovations in dynamic multiscale modulation, moving from recurrent and convolutional neural networks-with fixed scalings-to attention, transformers, dynamic convolutions, and consciousness priors-which modulate scale to input and increase scale breadth. The use and development of these multiscale innovations in robotic agents, game AI, and natural language processing (NLP) are pushing the boundaries of AI achievements. By juxtaposing biological and artificial intelligence, the present work underscores the critical importance of multiscale processing to general intelligence, as well as highlighting innovations and differences between the future of biological and artificial intelligence.

Is It Possible to Improve Working Memory With Prefrontal tDCS? Bridging Currents to Working Memory Models

A great deal of research has been performed with the promise of improving such critical cognitive functions as working memory (WM), with transcranial direct current stimulation (tDCS), a well-tolerated, inexpensive, easy-to-use intervention. Under the assumption that by delivering currents through electrodes placed in suitable locations on the scalp, it is possible to increase prefrontal cortex excitability and therefore improve WM. A growing number of studies have led to mixed results, leading to the realization that such oversimplified assumptions need revision. Models spanning currents to behavior have been advocated in order to reconcile and inform neurostimulation investigations. We articulate such multilevel exploration to tDCS/WM by briefly reviewing critical aspects at each level of analysis but focusing on the circuit level and how available biophysical WM models could inform tDCS. Indeed, such models should replace vague reference to cortical excitability changes with relevant tDCS net effects affecting neural computation and behavior in a more predictable manner. We will refer to emerging WM models and explore to what extent the general concept of excitation-inhibition (E/I) balance is a meaningful intermediate level of analysis, its relationship with gamma oscillatory activity, and the extent to which it can index tDCS effects. We will highlight some predictions that appear consistent with empirical evidence – such as non-linearities and trait dependency of effects and possibly a preferential effect on WM control functions – as well as limitations that appear related to the dynamical aspects of coding by persistent activity.

Constructing and Forgetting Temporal Context in the Human Cerebral Cortex

How does information from seconds earlier affect neocortical responses to new input? We found that when two groups of participants heard the same sentence in a narrative, preceded by different contexts, the neural responses of each group were initially different but gradually fell into alignment. We observed a hierarchical gradient: sensory cortices aligned most quickly, followed by mid-level regions, while some higher-order cortical regions took more than 10 seconds to align. What computations explain this hierarchical temporal organization? Linear integration models predict that regions that are slower to integrate new information should also be slower to forget old information. However, we found that higher-order regions could rapidly forget prior context. The data from the cortical hierarchy were instead captured by a model in which each region maintains a temporal context representation that is nonlinearly integrated with input at each moment, and this integration is gated by local prediction error.

The Value of Persistent Value

In this issue of Neuron, Bari et al. (2019) show that neurons in medial frontal cortex, but not a nearby premotor area, encode the relative value of available options with long-lasting persistent activity states during naturalistic foraging. These long-lasting activity states serve to preferentially guide choices to more likely rewarded options.

Reevaluating the Role of Persistent Neural Activity in Short-Term Memory

A traditional view of short-term working memory (STM) is that task-relevant information is maintained 'online' in persistent spiking activity. However, recent experimental and modeling studies have begun to question this long-held belief. In this review, we discuss new evidence demonstrating that information can be 'silently' maintained via short-term synaptic plasticity (STSP) without the need for persistent activity. We discuss how the neural mechanisms underlying STM are inextricably linked with the cognitive demands of the task, such that the passive maintenance and the active manipulation of information are subserved differently in the brain. Together, these recent findings point towards a more nuanced view of STM in which multiple substrates work in concert to support our ability to temporarily maintain and manipulate information.

Drifting codes within a stable coding scheme for working memory

Working memory (WM) is important to maintain information over short time periods to provide some stability in a constantly changing environment. However, brain activity is inherently dynamic, raising a challenge for maintaining stable mental states. To investigate the relationship between WM stability and neural dynamics, we used electroencephalography to measure the neural response to impulse stimuli during a WM delay. Multivariate pattern analysis revealed representations were both stable and dynamic: there was a clear difference in neural states between time-specific impulse responses, reflecting dynamic changes, yet the coding scheme for memorised orientations was stable. This suggests that a stable subcomponent in WM enables stable maintenance within a dynamic system. A stable coding scheme simplifies readout for WM-guided behaviour, whereas the low-dimensional dynamic component could provide additional temporal information. Despite having a stable subspace, WM is clearly not perfect-memory performance still degrades over time. Indeed, we find that even within the stable coding scheme, memories drift during maintenance. When averaged across trials, such drift contributes to the width of the error distribution.

Transformation of speech sequences in human sensorimotor circuits

After we listen to a series of words, we can silently replay them in our mind. Does this mental replay involve a reactivation of our original perceptual dynamics? We recorded electrocorticographic (ECoG) activity across the lateral cerebral cortex as people heard and then mentally rehearsed spoken sentences. For each region, we tested whether silent rehearsal of sentences involved reactivation of sentence-specific representations established during perception or transformation to a distinct representation. In sensorimotor and premotor cortex, we observed reliable and temporally precise responses to speech; these patterns transformed to distinct sentence-specific representations during mental rehearsal. In contrast, we observed less reliable and less temporally precise responses in prefrontal and temporoparietal cortex; these higher-order representations, which were sensitive to sentence semantics, were shared across perception and rehearsal of the same sentence. The mental rehearsal of natural speech involves the transformation of stimulus-locked speech representations in sensorimotor and premotor cortex, combined with diffuse reactivation of higher-order semantic representations.

Simple framework for constructing functional spiking recurrent neural networks

Cortical microcircuits exhibit complex recurrent architectures that possess dynamically rich properties. The neurons that make up these microcircuits communicate mainly via discrete spikes, and it is not clear how spikes give rise to dynamics that can be used to perform computationally challenging tasks. In contrast, continuous models of rate-coding neurons can be trained to perform complex tasks. Here, we present a simple framework to construct biologically realistic spiking recurrent neural networks (RNNs) capable of learning a wide range of tasks. Our framework involves training a continuous-variable rate RNN with important biophysical constraints and transferring the learned dynamics and constraints to a spiking RNN in a one-to-one manner. The proposed framework introduces only 1 additional parameter to establish the equivalence between rate and spiking RNN models. We also study other model parameters related to the rate and spiking networks to optimize the one-to-one mapping. By establishing a close relationship between rate and spiking models, we demonstrate that spiking RNNs could be constructed to achieve similar performance as their counterpart continuous rate networks.

Simple framework for constructing functional spiking recurrent neural networks

Cortical microcircuits exhibit complex recurrent architectures that possess dynamically rich properties. The neurons that make up these microcircuits communicate mainly via discrete spikes, and it is not clear how spikes give rise to dynamics that can be used to perform computationally challenging tasks. In contrast, continuous models of rate-coding neurons can be trained to perform complex tasks. Here, we present a simple framework to construct biologically realistic spiking recurrent neural networks (RNNs) capable of learning a wide range of tasks. Our framework involves training a continuous-variable rate RNN with important biophysical constraints and transferring the learned dynamics and constraints to a spiking RNN in a one-to-one manner. The proposed framework introduces only 1 additional parameter to establish the equivalence between rate and spiking RNN models. We also study other model parameters related to the rate and spiking networks to optimize the one-to-one mapping. By establishing a close relationship between rate and spiking models, we demonstrate that spiking RNNs could be constructed to achieve similar performance as their counterpart continuous rate networks.

Premembering Experience: A Hierarchy of Time-Scales for Proactive Attention

Memories are about the past, but they serve the future. Memory research often emphasizes the former aspect: focusing on the functions that re-constitute (re-member) experience and elucidating the various types of memories and their interrelations, timescales, and neural bases. Here we highlight the prospective nature of memory in guiding selective attention, focusing on functions that use previous experience to anticipate the relevant events about to unfold-to "premember" experience. Memories of various types and timescales play a fundamental role in guiding perception and performance adaptively, proactively, and dynamically. Consonant with this perspective, memories are often recorded according to expected future demands. Using working memory as an example, we consider how mnemonic content is selected and represented for future use. This perspective moves away from the traditional representational account of memory toward a functional account in which forward-looking memory traces are informationally and computationally tuned for interacting with incoming sensory signals to guide adaptive behavior.

Immediate versus delayed control demands elicit distinct mechanisms for instantiating proactive control

Cognitive control is critical for dynamically guiding goal-directed behavior, particularly when applying preparatory, or proactive, control processes. However, it is unknown how proactive control is modulated by timing demands. This study investigated how timing demands may instantiate distinct neural processes and contribute to the use of different types of proactive control. In two experiments, healthy young adults performed the AX-Continuous Performance Task (AX-CPT) or Dot Pattern Expectancy (DPX) task. The delay between informative cue and test probe was manipulated by block to be short (1s) or long (~3s). We hypothesized that short cue-probe delays would rely more on a rapid goal updating process (akin to task-switching), whereas long cue-probe delays would utilize more of an active maintenance process (akin to working memory). Short delay lengths were associated with specific impairments in rare probe accuracy. EEG responses to control-demanding cues revealed delay-specific neural signatures, which replicated across studies. In the short delay condition, EEG activities associated with task-switching were specifically enhanced, including increased early anterior positivity ERP amplitude (accompanying greater mid-frontal theta power) and a larger late differential switch positivity. In the long delay condition, we observed study-specific sustained increases in ERP amplitude following control-demanding cues, which may be suggestive of active maintenance. Collectively, these findings suggest that timing demands may instantiate distinct proactive control processes. These findings suggest a reevaluation of AX-CPT and DPX as pure assessments of working memory and highlight the need to understand how presumably benign task parameters, such as cue-probe delay length, significantly alter cognitive control.

Working Memory 2.0

Working memory is the fundamental function by which we break free from reflexive input-output reactions to gain control over our own thoughts. It has two types of mechanisms: online maintenance of information and its volitional or executive control. Classic models proposed persistent spiking for maintenance but have not explicitly addressed executive control. We review recent theoretical and empirical studies that suggest updates and additions to the classic model. Synaptic weight changes between sparse bursts of spiking strengthen working memory maintenance. Executive control acts via interplay between network oscillations in gamma (30-100 Hz) in superficial cortical layers (layers 2 and 3) and alpha and beta (10-30 Hz) in deep cortical layers (layers 5 and 6). Deep-layer alpha and beta are associated with top-down information and inhibition. It regulates the flow of bottom-up sensory information associated with superficial layer gamma. We propose that interactions between different rhythms in distinct cortical layers underlie working memory maintenance and its volitional control.