Dynamical latent state computation in the male macaque posterior parietal cortex.

Success in many real-world tasks depends on our ability to dynamically track hidden states of the world. We hypothesized that neural populations estimate these states by processing sensory history through recurrent interactions which reflect the internal model of the world. To test this, we recorded brain activity in posterior parietal cortex (PPC) of monkeys navigating by optic flow to a hidden target location within a virtual environment, without explicit position cues. In addition to sequential neural dynamics and strong interneuronal interactions, we found that the hidden state - monkey's displacement from the goal - was encoded in single neurons, and could be dynamically decoded from population activity. The decoded estimates predicted navigation performance on individual trials. Task manipulations that perturbed the world model induced substantial changes in neural interactions, and modified the neural representation of the hidden state, while representations of sensory and motor variables remained stable. The findings were recapitulated by a task-optimized recurrent neural network model, suggesting that task demands shape the neural interactions in PPC, leading them to embody a world model that consolidates information and tracks task-relevant hidden states.

Coding of latent variables in sensory, parietal, and frontal cortices during closed-loop virtual navigation.

We do not understand how neural nodes operate and coordinate within the recurrent action-perception loops that characterize naturalistic self-environment interactions. Here, we record single-unit spiking activity and local field potentials (LFPs) simultaneously from the dorsomedial superior temporal area (MSTd), parietal area 7a, and dorsolateral prefrontal cortex (dlPFC) as monkeys navigate in virtual reality to 'catch fireflies'. This task requires animals to actively sample from a closed-loop virtual environment while concurrently computing continuous latent variables: (i) the distance and angle travelled (i.e., path integration) and (ii) the distance and angle to a memorized firefly location (i.e., a hidden spatial goal). We observed a patterned mixed selectivity, with the prefrontal cortex most prominently coding for latent variables, parietal cortex coding for sensorimotor variables, and MSTd most often coding for eye movements. However, even the traditionally considered sensory area (i.e., MSTd) tracked latent variables, demonstrating path integration and vector coding of hidden spatial goals. Further, global encoding profiles and unit-to-unit coupling (i.e., noise correlations) suggested a functional subnetwork composed by MSTd and dlPFC, and not between these and 7a, as anatomy would suggest. We show that the greater the unit-to-unit coupling between MSTd and dlPFC, the more the animals' gaze position was indicative of the ongoing location of the hidden spatial goal. We suggest this MSTd-dlPFC subnetwork reflects the monkeys' natural and adaptive task strategy wherein they continuously gaze toward the location of the (invisible) target. Together, these results highlight the distributed nature of neural coding during closed action-perception loops and suggest that fine-grain functional subnetworks may be dynamically established to subserve (embodied) task strategies.

Purkinje Cell Activity in the Medial and Lateral Cerebellum During Suppression of Voluntary Eye Movements in Rhesus Macaques.

Volitional suppression of responses to distracting external stimuli enables us to achieve our goals. This volitional inhibition of a specific behavior is supposed to be mainly mediated by the cerebral cortex. However, recent evidence supports the involvement of the cerebellum in this process. It is currently not known whether different parts of the cerebellar cortex play differential or synergistic roles in the planning and execution of this behavior. Here, we measured Purkinje cell (PC) responses in the medial and lateral cerebellum in two rhesus macaques during pro- and anti-saccade tasks. During an antisaccade trial, non-human primates (NHPs) were instructed to make a saccadic eye movement away from a target, rather than toward it, as in prosaccade trials. Our data show that the cerebellum plays an important role not only during the execution of the saccades but also during the volitional inhibition of eye movements toward the target. Simple spike (SS) modulation during the instruction and execution periods of pro- and anti-saccades was prominent in PCs of both the medial and lateral cerebellum. However, only the SS activity in the lateral cerebellar cortex contained information about stimulus identity and showed a strong reciprocal interaction with complex spikes (CSs). Moreover, the SS activity of different PC groups modulated bidirectionally in both of regions, but the PCs that showed facilitating and suppressive activity were predominantly associated with instruction and execution, respectively. These findings show that different cerebellar regions and PC groups contribute to goal-directed behavior and volitional inhibition, but with different propensities, highlighting the rich repertoire of the cerebellar control in executive functions.

Sensory Evidence Accumulation Using Optic Flow in a Naturalistic Navigation Task.

Sensory evidence accumulation is considered a hallmark of decision-making in noisy environments. Integration of sensory inputs has been traditionally studied using passive stimuli, segregating perception from action. Lessons learned from this approach, however, may not generalize to ethological behaviors like navigation, where there is an active interplay between perception and action. We designed a sensory-based sequential decision task in virtual reality in which humans and monkeys navigated to a memorized location by integrating optic flow generated by their own joystick movements. A major challenge in such closed-loop tasks is that subjects' actions will determine future sensory input, causing ambiguity about whether they rely on sensory input rather than expectations based solely on a learned model of the dynamics. To test whether subjects integrated optic flow over time, we used three independent experimental manipulations, unpredictable optic flow perturbations, which pushed subjects off their trajectory; gain manipulation of the joystick controller, which changed the consequences of actions; and manipulation of the optic flow density, which changed the information borne by sensory evidence. Our results suggest that both macaques (male) and humans (female/male) relied heavily on optic flow, thereby demonstrating a critical role for sensory evidence accumulation during naturalistic action-perception closed-loop tasks.SIGNIFICANCE STATEMENT The temporal integration of evidence is a fundamental component of mammalian intelligence. Yet, it has traditionally been studied using experimental paradigms that fail to capture the closed-loop interaction between actions and sensations inherent in real-world continuous behaviors. These conventional paradigms use binary decision tasks and passive stimuli with statistics that remain stationary over time. Instead, we developed a naturalistic visuomotor visual navigation paradigm that mimics the causal structure of real-world sensorimotor interactions and probed the extent to which participants integrate sensory evidence by adding task manipulations that reveal complementary aspects of the computation.

Spatial modulation of hippocampal activity in freely moving macaques.

The hippocampal formation is linked to spatial navigation, but there is little corroboration from freely moving primates with concurrent monitoring of head and gaze stances. We recorded neural activity across hippocampal regions in rhesus macaques during free foraging in an open environment while tracking their head and eye. Theta activity was intermittently present at movement onset and modulated by saccades. Many neurons were phase-locked to theta, with few showing phase precession. Most neurons encoded a mixture of spatial variables beyond place and grid tuning. Spatial representations were dominated by facing location and allocentric direction, mostly in head, rather than gaze, coordinates. Importantly, eye movements strongly modulated neural activity in all regions. These findings reveal that the macaque hippocampal formation represents three-dimensional (3D) space using a multiplexed code, with head orientation and eye movement properties being dominant during free exploration.

B.2.16 is a non-lethal modifier of the Dark82 mosaic eye phenotype in Drosophila melanogaster.

Genetic screens have been used to identify genes involved in the regulation of different biological processes. We identified growth mutants in a Flp/FRT screen using the Drosophila melanogaster eye to identify conditional regulators of cell growth and cell division. One mutant identified from this screen, B.2.16, was mapped and characterized by researchers in undergraduate genetics labs as part of the Fly-CURE. We find that B.2.16 is a non-lethal genetic modifier of the Dark82 mosaic eye phenotype.

Supporting generalization in non-human primate behavior by tapping into structural knowledge: Examples from sensorimotor mappings, inference, and decision-making.

The complex behaviors we ultimately wish to understand are far from those currently used in systems neuroscience laboratories. A salient difference are the closed loops between action and perception prominently present in natural but not laboratory behaviors. The framework of reinforcement learning and control naturally wades across action and perception, and thus is poised to inform the neurosciences of tomorrow, not only from a data analyses and modeling framework, but also in guiding experimental design. We argue that this theoretical framework emphasizes active sensing, dynamical planning, and the leveraging of structural regularities as key operations for intelligent behavior within uncertain, time-varying environments. Similarly, we argue that we may study natural task strategies and their neural circuits without over-training animals when the tasks we use tap into our animal's structural knowledge. As proof-of-principle, we teach animals to navigate through a virtual environment - i.e., explore a well-defined and repetitive structure governed by the laws of physics - using a joystick. Once these animals have learned to 'drive', without further training they naturally (i) show zero- or one-shot learning of novel sensorimotor contingencies, (ii) infer the evolving path of dynamically changing latent variables, and (iii) make decisions consistent with maximizing reward rate. Such task designs allow for the study of flexible and generalizable, yet controlled, behaviors. In turn, they allow for the exploitation of pillars of intelligence - flexibility, prediction, and generalization -, properties whose neural underpinning have remained elusive.

Tracking the Mind's Eye: Primate Gaze Behavior during Virtual Visuomotor Navigation Reflects Belief Dynamics.

To take the best actions, we often need to maintain and update beliefs about variables that cannot be directly observed. To understand the principles underlying such belief updates, we need tools to uncover subjects' belief dynamics from natural behavior. We tested whether eye movements could be used to infer subjects' beliefs about latent variables using a naturalistic navigation task. Humans and monkeys navigated to a remembered goal location in a virtual environment that provided optic flow but lacked explicit position cues. We observed eye movements that appeared to continuously track the goal location even when no visible target was present there. Accurate goal tracking was associated with improved task performance, and inhibiting eye movements in humans impaired navigation precision. These results suggest that gaze dynamics play a key role in action selection during challenging visuomotor behaviors and may possibly serve as a window into the subject's dynamically evolving internal beliefs.

Visual and Vestibular Selectivity for Self-Motion in Macaque Posterior Parietal Area 7a.

We examined the responses of neurons in posterior parietal area 7a to passive rotational and translational self-motion stimuli, while systematically varying the speed of visually simulated (optic flow cues) or actual (vestibular cues) self-motion. Contrary to a general belief that responses in area 7a are predominantly visual, we found evidence for a vestibular dominance in self-motion processing. Only a small fraction of neurons showed multisensory convergence of visual/vestibular and linear/angular self-motion cues. These findings suggest possibly independent neuronal population codes for visual versus vestibular and linear versus angular self-motion. Neural responses scaled with self-motion magnitude (i.e., speed) but temporal dynamics were diverse across the population. Analyses of laminar recordings showed a strong distance-dependent decrease for correlations in stimulus-induced (signal correlation) and stimulus-independent (noise correlation) components of spike-count variability, supporting the notion that neurons are spatially clustered with respect to their sensory representation of motion. Single-unit and multiunit response patterns were also correlated, but no other systematic dependencies on cortical layers or columns were observed. These findings describe a likely independent multimodal neural code for linear and angular self-motion in a posterior parietal area of the macaque brain that is connected to the hippocampal formation.

Cerebellar tDCS Does Not Enhance Performance in an Implicit Categorization Learning Task.

Background: Transcranial Direct Current Stimulation (tDCS) is a form of non-invasive electrical stimulation that changes neuronal excitability in a polarity and site-specific manner. In cognitive tasks related to prefrontal and cerebellar learning, cortical tDCS arguably facilitates learning, but the few studies investigating cerebellar tDCS, however, are inconsistent. Objective: We investigate the effect of cerebellar tDCS on performance of an implicit categorization learning task. Methods: Forty participants performed a computerized version of an implicit categorization learning task where squares had to be sorted into two categories, according to an unknown but fixed rule that integrated both the size and luminance of the square. Participants did one round of categorization to familiarize themselves with the task and to provide a baseline of performance. After that, 20 participants received anodal tDCS (20 min, 1.5 mA) over the right cerebellum, and 19 participants received sham stimulation and simultaneously started a second session of the categorization task using a new rule. Results: As expected, subjects performed better in the second session than in the first, baseline session, showing increased accuracy scores and reduced reaction times. Over trials, participants learned the categorization rule, improving their accuracy and reaction times. However, we observed no effect of anodal tDCS stimulation on overall performance or on learning, compared to sham stimulation. Conclusion: These results suggest that cerebellar tDCS does not modulate performance and learning on an implicit categorization task.

Cerebellar transcranial direct current stimulation effects on saccade adaptation.

Saccade adaptation is a cerebellar-mediated type of motor learning in which the oculomotor system is exposed to repetitive errors. Different types of saccade adaptations are thought to involve distinct underlying cerebellar mechanisms. Transcranial direct current stimulation (tDCS) induces changes in neuronal excitability in a polarity-specific manner and offers a modulatory, noninvasive, functional insight into the learning aspects of different brain regions. We aimed to modulate the cerebellar influence on saccade gains during adaptation using tDCS. Subjects performed an inward (n = 10) or outward (n = 10) saccade adaptation experiment (25% intrasaccadic target step) while receiving 1.5 mA of anodal cerebellar tDCS delivered by a small contact electrode. Compared to sham stimulation, tDCS increased learning of saccadic inward adaptation but did not affect learning of outward adaptation. This may imply that plasticity mechanisms in the cerebellum are different between inward and outward adaptation. TDCS could have influenced specific cerebellar areas that contribute to inward but not outward adaptation. We conclude that tDCS can be used as a neuromodulatory technique to alter cerebellar oculomotor output, arguably by engaging wider cerebellar areas and increasing the available resources for learning.