This letter offers a computational account of Pavlovian conditioning in the cerebellum based on active inference and predictive coding. Using eyeblink conditioning as a canonical paradigm, we formulate a minimal generative model that can account for spontaneous blinking, startle responses, and (delay or trace) conditioning. We then establish the face validity of the model using simulated responses to unconditioned and conditioned stimuli to reproduce the sorts of behavior that are observed empirically. The scheme’s anatomical validity is then addressed by associating variables in the predictive coding scheme with nuclei and neuronal populations to match the (extrinsic and intrinsic) connectivity of the cerebellar (eyeblink conditioning) system. Finally, we try to establish predictive validity by reproducing selective failures of delay conditioning, trace conditioning, and extinction using (simulated and reversible) focal lesions. Although rather metaphorical, the ensuing scheme can account for a remarkable range of anatomical and neurophysiological aspects of cerebellar circuitry—and the specificity of lesion-deficit mappings that have been established experimentally. From a computational perspective, this work shows how conditioning or learning can be formulated in terms of minimizing variational free energy (or maximizing Bayesian model evidence) using exactly the same principles that underlie predictive coding in perception.

Background—The cerebellum has been recently suggested as an important player in the addiction brain circuit. Cannabis is one of the most used drugs worldwide, and its long-term effects on the central nervous system are not fully understood. No valid clinical evaluations of cannabis impact on the brain are available today. The cerebellum is expected to be one of the brain structures that are highly affected by prolonged exposure to cannabis, due to its high density in endocannabinoid receptors. We aim to use a motor adaptation paradigm to indirectly assess cerebellar function in chronic cannabis users (CCUs). Methods—We used a visuomotor rotation (VMR) task that probes a putatively-cerebellar implicit motor adaptation process together with the learning and execution of an explicit aiming rule. We conducted a case-control study, recruiting 18 CCUs and 18 age-matched healthy controls. Our main measure was the angular aiming error. Results—Our results show that CCUs have impaired implicit motor adaptation, as they showed a smaller rate of adaptation compared with healthy controls (drift rate: 19.3 +/− 6.8° vs. 27.4 +/− 11.6°; t(26) = −2.1, p = 0.048, Cohen’s d = −0.8, 95% CI = (−1.7, −0.15)). Conclusions—We suggest that a visuomotor rotation task might be the first step towards developing a useful tool for the detection of alterations in implicit learning among cannabis users.

In this perspective article, we show that a morphospace, based on information-theoretic measures, can be a useful construct for comparing biological agents with artificial intelligence (AI) systems. The axes of this space label three kinds of complexity: (i) autonomic, (ii) computational and (iii) social complexity. On this space, we map biological agents such as bacteria, bees, C. elegans, primates and humans; as well as AI technologies such as deep neural networks, multi-agent bots, social robots, Siri and Watson. A complexity-based conceptualization provides a useful framework for identifying defining features and classes of conscious and intelligent systems. Starting with cognitive and clinical metrics of consciousness that assess awareness and wakefulness, we ask how AI and synthetically engineered life-forms would measure on homologous metrics. We argue that awareness and wakefulness stem from computational and autonomic complexity. Furthermore, tapping insights from cognitive robotics, we examine the functional role of consciousness in the context of evolutionary games. This points to a third kind of complexity for describing consciousness, namely, social complexity. Based on these metrics, our morphospace suggests the possibility of additional types of consciousness other than biological; namely, synthetic, group-based and simulated. This space provides a common conceptual framework for comparing traits and highlighting design principles of minds and machines.

How does our motor system solve the problem of anticipatory control in spite of a wide spectrum of response dynamics from different musculo-skeletal systems, transport delays as well as response latencies throughout the central nervous system? To a great extent, our highly-skilled motor responses are a result of a reactive feedback system, originating in the brain-stem and spinal cord, combined with a feed-forward anticipatory system, that is adaptively fine-tuned by sensory experience and originates in the cerebellum. Based on that interaction we design the counterfactual predictive control (CFPC) architecture, an anticipatory adaptive motor control scheme, in which a feed-forward module, based on the cerebellum, steers an error feedback controller with counterfactual error signals. Those are signals that trigger reactions as actual errors would, but that do not code for any current of forthcoming errors. In order to determine the optimal learning strategy, we derive a novel learning rule for the feed-forward module that involves an eligibility trace and operates at the synaptic level. In particular, our eligibility trace provides a mechanism beyond co-incidence detection in that it convolves a history of prior synaptic inputs with error signals. In the context of cerebellar physiology, this solution implies that Purkinje cell synapses should generate eligibility traces using a forward model of the system being controlled. From an engineering perspective, CFPC provides a general-purpose anticipatory control architecture equipped with a learning rule that exploits the full dynamics of the closed-loop system.

Humans display anticipatory motor responses to minimize the adverse effects of predictable perturbations. A widely accepted explanation for this behavior relies on the notion of an inverse model that, learning from motor errors, anticipates corrective responses. Here, we propose and validate the alternative hypothesis that anticipatory control can be realized through a cascade of purely sensory predictions that drive the motor system, reflecting the causal sequence of the perceptual events preceding the error. We compare both hypotheses in a simulated anticipatory postural adjustment task. We observe that adaptation in the sensory domain, but not in the motor one, supports the robust and generalizable anticipatory control characteristic of biological systems. Our proposal unites the neurobiology of the cerebellum with the theory of active inference and provides a concrete implementation of its core tenets with great relevance both to our understanding of biological control systems and, possibly, to their emulation in complex artefacts.

How does our nervous system successfully acquire feedback control strategies in spite of a wide spectrum of response dynamics from different musculo-skeletal systems? The cerebellum is a crucial brain structure in enabling precise motor control in animals. Recent advances suggest that synaptic plasticity of cerebellar Purkinje cells involves molecular mechanisms that mimic the dynamics of the efferent motor system that they control allowing them to match the timing of their learning rule to behavior. Counter-Factual Predictive Control (CFPC) is a cerebellum-based feed-forward control scheme that exploits that principle for acquiring anticipatory actions. CFPC extends the classical Widrow-Hoff/Least Mean Squares by inserting a forward model of the downstream closed-loop system in its learning rule. Here we apply that same insight to the problem of learning the gains of a feedback controller. To that end, we frame a Model-Reference Adaptive Control (MRAC) problem and derive an adaptive control scheme treating the gains of a feedback controller as if they were the weights of an adaptive linear unit. Our results demonstrate that rather than being exclusively confined to cerebellar learning, the approach of controlling plasticity with a forward model of the subsystem controlled, an approach that we term as Model-Enhanced Least Mean Squares (ME-LMS), can provide a solution to wide set of adaptive control problems.