Biology subjects

de la Rocha, J.

Publications and source records attributed to de la Rocha, J..

15 recordsLinked to original sources

bioRxiv2026

Interplay of pupil-linked arousal and cortical decision computations in mice

Influential frameworks hold that brainstem neuromodulatory systems are rapidly recruited in a top-down manner during decision-making, adapting brain-wide activity and cognitive behavior in real time. Here, we examined the relationship between decision-related frontal cortical activity and pupil responses (a peripheral proxy of neuromodulatory activity) in mice accumulating fluctuating auditory evidence to choose among two options. We found two components of preparatory activity during decision formation, prior to motor action: a non-selective component shared between hemispheres and a choice-selective component that reflected decision uncertainty. Both components were mirrored in pupil responses during the task, which, in turn, scaled with decision uncertainty during decision formation and with uncertainty-modulated prediction errors after experiencing the outcome. Pupil responses also predicted an adjustment of choice behavior at the next trial. Our results point to a recurrent interplay between cortical decision computation and pupil-linked neuromodulation. HIGHLIGHTSO_LIMice switch between evidence integration and heuristic strategies during auditory-informed decisions C_LIO_LIPupil responses track decision uncertainty and uncertainty-modulated prediction errors C_LIO_LIDecision-related frontal cortical activity is coupled with pupil responses C_LIO_LIPupil responses predict an adjustment of future choice behavior C_LI IN SHORTMaheu et al. show that the phasic transitions in the arousal state mirror distinct components of cortical activity, including one directly involved in decision computations. In turn, these pupil-linked arousal responses predicted the strategy adopted at the next trial.

neuroscience↗

bioRxiv2026

History-dependent Biases in Perceptual Decisions Depend on NMDA Receptors

Perceptual decisions are governed not only by current sensory input but also by the history of previous stimuli, choices, or outcomes. Such history-dependent biases evolve across multiple timescales ranging from seconds to minutes. Synaptic plasticity is one possible mechanism mediating such effects. The by N-methyl-aspartate (NMDA) glutamate receptor is critical for synaptic plasticity, and indirect evidence points to its role in the generation of history-dependent biases. Here, we pharmacologically blocked NMDA receptors in healthy human participants performing a two-alternative visual decision-making task. Participants choices were biased by the accumulation of fluctuating sensory evidence within each trial as well as the accumulation of across-trial sequences of stimuli, which enabled them to adapt to changing stimulus statistics across blocks. NMDA receptor blockade reduced the task-adaptive across trials, history-dependent biases. In contrast, enhancing GABA-A receptor activity in the same participants increased the transiency of the within-trial evidence accumulation dynamics. Our findings provide direct evidence for the importance of NMDA receptor function in history-dependent biases in the human brain.

neuroscience↗

bioRxiv2026

The Training Village: an open platform for continuous testing of rodents in cognitive tasks

The development of new methods for dissecting neural circuits in rodents promises to transform the study of higher-order cognitive functions. However, the adaptation of complex cognitive tasks from humans to rodents remains very limited as these tasks require long periods of animal training with significant researcher expertise and effort. Multi-box setups can mitigate the problem but remain labor-intensive and costly to maintain. Newly developed home cage training systems, while minimizing costs, are limited to specific conditions and tasks. Here, we present the Training Village (TV), an open-source, affordable, and fully automated system for continuous rodent training. In the TV, a group of rodents lives in enriched arenas while individually accessing an operant box at any time. The system provides personalized training, continuous monitoring of task performance and home cage activity, and real-time remote supervision. Its graphical interface and modular design make it user-friendly and easy to integrate with other behavioral systems running a wide range of cognitive paradigms. We validated the TV across multiple tasks and cohorts of mice and rats, demonstrating efficient operant box usage, stable long-term task engagement, and its potential to link home cage behavior with task performance. Overall, the TV is an easy-to-adopt platform that can greatly accelerate brain research by fully automating rodent training in complex cognitive tasks.

animal behavior and cognition↗

bioRxiv2025

Competing Neural Decision Variables in Human Frontal Cortex Shape Decision Confidence

Theories of the neural basis of decision-making postulate a competition between choice-selective populations of neurons that are distributed across many brain areas. We hypothesized that agents sense of confidence in a choice is shaped by the consensus among these distributed decision signals. To test this idea, we simultaneously tracked the dynamics of choice-selective signals within each of many cortical areas via magnetoencephalography, while human participants formed difficult decisions. Signals within premotor cortex that supported the choice option selected on a given trial, or the alternative, exhibited a signature of intra-cortical competition. Both competing neural signals made opposing and unique contributions to confidence. Critically, the number and strength of choice-selective signals in associative cortical areas supporting the selected option on each trial made an additional contribution to confidence. We conclude that confidence emerges from cortex-wide, competitive dynamics, through a mechanism that monitors the consensus across many local decision modules.

neuroscience↗

bioRxiv2024

Episodic recruitment of attractor dynamics in frontal cortex reveals distinct mechanisms for forgetting and lack of cognitive control in short-term memory

Short-term memory (STM) is prone to failure, especially during prolonged memory maintenance or under limited cognitive control. Despite predictive mechanistic frameworks based on persistent neural activity and attractor states, a direct assessment of network dynamics during multifactorial STM failure is still missing. We addressed this in a delayed-response task where mice maintained a prospective response during a long variable delay. Mice behavior episodically switched between a task-engaged state described by an attractor model, and a task-disengaged state purely determined by previous choices. During task engagement, the anterolateral motor cortex (ALM) showed delay persistent activity stably encoding correct choices, whereas the encoding reversed during the delay in error trials. In contrast, in task-disengaged phases ALM showed no clear traces of attractor dynamics and instead exhibited enhanced synchrony at [~] 4-5Hz. Thus, ALM switches between distinct error-generating dynamics: in control-capable trials, transitions between memory attractors cause forgetting errors, whereas non-memory errors are caused by the dissociation of ALM during the mnemonic period reflecting the lack of cognitive control.

neuroscience↗

bioRxiv2023

Fast and tight control of motor vigor by accumulated decision evidence

Acting in the natural world requires not only deciding among multiple options but also converting decisions into motor commands. How the dynamics of decision formation influence the fine kinematics of response movement remains, however, poorly understood. Here we investigate how the accumulation of decision evidence shapes the response orienting trajectories in a task where freely-moving rats combine prior expectations and auditory information to select between two possible options. Response trajectories and their motor vigor are initially determined by the prior. Rats movements then incorporate sensory information as early as 60 ms after stimulus onset by accelerating or slowing depending on how much the stimulus supports their initial choice. When the stimulus evidence is in strong contradiction, rats change their mind and reverse their initial trajectory. Human subjects performing an equivalent task display a remarkably similar behavior. We encapsulate these results in a computational model that, by mapping the decision variable onto the movement kinematics at discrete time points, captures subjects choices, trajectories and changes of mind. Our results show that motor responses are not ballistic. Instead, they are systematically and rapidly updated, as they smoothly unfold over time, by the parallel dynamics of the underlying decision process.

neuroscience↗

bioRxiv2022

Temporal integration is a robust feature of perceptual decisions

Making informed decisions in noisy environments requires integrating sensory information over time. However, recent work has suggested that it may be difficult to determine whether an animals decision-making strategy relies on evidence integration or not. In particular, strategies based on extrema-detection or random snapshots of the evidence stream may be difficult or even impossible to distinguish from classic evidence integration. Moreover, such non-integration strategies might be surprisingly common in experiments that aimed to study decisions based on integration. To determine whether temporal integration is central to perceptual decision making, we developed a new model-based approach for comparing temporal integration against alternative "non-integration" strategies for tasks in which the sensory signal is composed of discrete stimulus samples. We applied these methods to behavioral data from monkeys, rats, and humans performing a variety of sensory decision-making tasks. In all species and tasks, we found converging evidence in favor of temporal integration. First, in all observers across studies, the integration model better accounted for standard behavioral statistics such as psychometric curves and psychophysical kernels. Second, we found that sensory samples with large evidence do not contribute disproportionately to subject choices, as predicted by an extrema-detection strategy. Finally, we provide a direct confirmation of temporal integration by showing that the sum of both early and late evidence contributed to observer decisions. Overall, our results provide experimental evidence suggesting that temporal integration is an ubiquitous feature in mammalian perceptual decision-making. Our study also highlights the benefits of using experimental paradigms where the temporal stream of sensory evidence is controlled explicitly by the experimenter, and known precisely by the analyst, to characterize the temporal properties of the decision process.

neuroscience↗

bioRxiv2022

Isolating perceptual biases caused by trial history during auditory categorization

Just as most experiences have their origin in our perceptions, our perceptions can also be fundamentally shaped by our experiences. However, isolating which events in the recent past can impact perceptual judgments remains a difficult question, partly because post-perceptual processes can also introduce strong history dependencies. Two mechanisms have been hypothesized to specifically modulate perception: 1) the repulsive influence caused by previous stimuli and generally labeled as aftereffects, and 2) the modulation caused by stimulus predictions based on learned temporal regularities of the sensory environment, a key assumption in the predictive coding framework. Here, we ask whether these two mechanisms do indeed bias perception by training rats in an auditory task featuring serial correlations along the sequence of stimuli. We develop a detailed behavioral model that isolates the repulsive aftereffect generated by previous stimuli and shows that this repulsion cannot be explained from an interaction between past and current stimuli, and that it is still present in catch trials lacking the current stimulus. Moreover, the model describes that the bias caused by the animals expectation, as they leverage the predictability of the stimulus sequence, is present in a foraging task without the sensory component but with the same serial correlations in the sequence of rewards. These results indicate that the aftereffect and the prediction biases observed during an auditory task can all be revealed in the absence of a sensory stimulus, putting into question their perceptual nature.

neuroscience↗

bioRxiv2021

Pre-training RNNs on ecologically relevant tasks explains sub-optimal behavioral reset

The strategies found by animals facing a new task are determined both by individual experience and by structural priors evolved to leverage the statistics of natural environments. Rats can quickly learn to capitalize on the trial sequence correlations of two-alternative forced choice (2AFC) tasks after correct trials, but consistently deviate from optimal behavior after error trials, when they waive the accumulated evidence. To understand this outcome-dependent gating, we first show that Recurrent Neural Networks (RNNs) trained in the same 2AFC task outperform rats as they can readily learn to use across-trial information both after correct and error trials. We hypothesize that, while RNNs can optimize their behavior in the 2AFC task without any a priori restrictions, rats strategy is constrained by a structural prior adapted to a natural environment in which rewarded and non-rewarded actions provide largely asymmetric information. When pre-training RNNs in a more ecological task with more than two possible choices, networks develop a strategy by which they gate off the across-trial evidence after errors, mimicking rats behavior. Population analyses show that the pre-trained networks form an accurate representation of the sequence statistics independently of the outcome in the previous trial. After error trials, gating is implemented by a change in the network dynamics which temporarily decouples the categorization of the stimulus from the across-trial accumulated evidence. Our results suggest that the suboptimal behavior observed in rats reflects the influence of a structural prior that reacts to errors by isolating the network decision dynamics from the context, ultimately constraining the performance in a 2AFC laboratory task.

neuroscience↗

bioRxiv2020

Flexible categorization in perceptual decision making

Perceptual decisions require the brain to make categorical choices based on accumulated sensory evidence. The underlying computations have been studied using either phenomenological drift diffusion models or neurobiological network models exhibiting winner-take-all attractor dynamics. Although both classes of models can account for a large body of experimental data, it remains unclear to what extent their dynamics are qualitatively equivalent. Here we show that, unlike the drift diffusion model, the attractor model can operate in different integration regimes: an increase in the stimulus fluctuations or the stimulus duration promotes transitions between decision-states leading to a crossover between weighting mostly early evidence (primacy regime) to weighting late evidence (recency regime). Between these two limiting cases, we found a novel regime, which we name flexible categorization, in which fluctuations are strong enough to reverse initial categorizations, but only if they are incorrect. This asymmetry in the reversing probability results in a non-monotonic psychometric curve, a novel and distinctive feature of the attractor model. Finally, we show psychophysical evidence for the crossover between integration regimes predicted by the attractor model and for the relevance of this new regime. Our findings point to correcting transitions as an important yet overlooked feature of perceptual decision making.

neuroscience↗

bioRxiv2019

Cortical state transitions and stimulus response evolve along stiff and sloppy parameter dimensions, respectively

Previous research showed that spontaneous neuronal activity presents sloppiness: the collective behavior is strongly determined by a small number of parameter combinations, defined as "stiff" dimensions, while it is insensitive to many others ("sloppy" dimensions). Here, we analyzed neural population activity from the auditory cortex of anesthetized rats while the brain spontaneously transited through different synchronized and desynchronized states and intermittently received sensory inputs. We showed that cortical state transitions were determined by changes in stiff parameters associated with the activity of a core of neurons with low responses to stimuli and high centrality within the observed network. In contrast, stimulus-evoked responses evolved along sloppy dimensions associated with the activity of neurons with low centrality and displaying large ongoing and stimulus-evoked fluctuations without affecting the integrity of the network. Our results shed light on the interplay among stability, flexibility, and responsiveness of neuronal collective dynamics during intrinsic and induced activity.

neuroscience↗

bioRxiv2019

Septal GABAergic Inputs to CA1 Govern Contextual Memory Retrieval

Whether projections from the medial septum regulate the function of the CA1 hippocampus in episodic memory retrieval is not known. Here we show that septal GABAergic inputs to CA1 promote contextual fear memory, blocking the activation of parvalbumin-rich interneurons to facilitate Erk/MAP-kinase signaling in pyramidal cells during retrieval. Thus, suppression of feed-forward inhibition onto CA1 by septal GABAergic neurons gates contextual fear behavior.

neuroscience↗

bioRxiv2019

High-throughput task to study memory recall during spatial navigation in rodents

Spatial navigation is the most frequently used behavioral paradigm to study hippocampal dependent memory formation in rodents. However, commonly used tasks can present some limitations: i) they are labor intensive, preventing the implementation of parallel testing for high-throughput experimentation; ii) yield a low number of repeated trials, curtailing the statistical power; iii) are hard to combine with neural recordings, because tethering sometimes interferes with behavior; iv) are not based on overt behavioral responses that can be precisely timed, making difficult the identification of the underlying neural events; v) produce a low spatial coverage, limiting the characterization of neuronal patterns related to spatial information. To circumvent these limitations, we developed a spatial memory task that required minimal human intervention, allowed simultaneous and unsupervised testing of several mice, and yielded a high number of recall trials per session (up to ~20). Moreover, because recall sessions could be repeated over many days, the task provided enough statistical power to characterize in detail the animals behavior during memory recall, even to quantify the decay in spatial accuracy of memories as they are stacked across days. In addition, the task is compatible with neural activity recordings. Together, these features make our task a valuable tool to start dissecting the neural circuit dynamics underlying spatial memory recall.

neuroscience↗

bioRxiv2018

Response outcomes gate the impact of expectations on perceptual decisions

Perceptual decisions are not only determined by current sensory information but are also influenced by recent experience. Can the magnitude of expectation-based decision biases be flexibly modulated based on the outcome of previous decisions? We trained rats in several two-alternative forced choice auditory tasks, where the probability to repeat the previous stimulus category was varied in blocks of trials. All rats capitalized on the serial correlations of the stimulus sequence by consistently exploiting a transition bias: a tendency to repeat or alternate their previous response using an internal trial-by-trial estimate of the sequence repeating probability. Surprisingly, this expectation bias was null in trials immediately following an error. The internal estimate however was not reset and it became effective again causing a bias after the next correct response. This ability to rapidly activate and inactivate the bias was captured by a non-linear generative model of rat behavior, whereby a reward-driven modulatory signal gated the use of the latent estimate of the environment statistics on the current decision. These results demonstrate that rats use behavioral outcomes to flexibly modulate how expectations derived from the recent history influence their decisions.

neuroscience↗

bioRxiv2016

UP-DOWN cortical dynamics reflect state transitions in a bistable balanced network

In the idling brain, neuronal circuits often exhibit transitions between periods of sustained firing (UP state) and quiescence (DOWN state). Although these dynamics occur across multiple areas and behavioral conditions, the underlying mechanisms remain unclear. Here we analyze spontaneous population activity from the somatosensory cortex of urethane-anesthetized rats. We find that UP and DOWN periods are variable (i.e. non-rhythmic) and that the population rate shows no significant decay during UP periods. We build a network model of excitatory (E) and inhibitory (I) neurons that exhibits a new bistability between a quiescent state and a balanced state of arbitrarily low rate. Fluctuating inputs trigger state transitions. Adaptation in E cells paradoxically causes a marginal decay of E-rate but a marked decay of I-rate, a signature of balanced bistability that we validate experimentally. Our findings provide evidence of a bistable balanced network that exhibits non-rhythmic state transitions when the brain rests.

neuroscience↗