Hirokazu Tanaka’s research while affiliated with Tokyo City University and other places

Cerebellum is a key-structure for the modulation of motor, cognitive, social and affective functions, contributing to automatic behaviours through interactions with the cerebral cortex, basal ganglia and spinal cord. The predictive mechanisms used by the cerebellum cover not only sensorimotor functions but also reward-related tasks. Cerebellar circuits appear to encode temporal difference error and reward prediction error. From a chemical standpoint, cerebellar catecholamines modulate the rate of cerebellar-based cognitive learning, and mediate cerebellar contributions during complex behaviours. Reward processing and its associated emotions are tuned by the cerebellum which operates as a controller of adaptive homeostatic processes based on interoceptive and exteroceptive inputs. Lobules VI-VII/areas of the vermis are candidate regions for the cortico-subcortical signaling pathways associated with loss aversion and reward sensitivity, together with other nodes of the limbic circuitry. There is growing evidence that the cerebellum works as a hub of regional dysconnectivity across all mood states and that mental disorders involve the cerebellar circuitry, including mood and addiction disorders, and impaired eating behaviors where the cerebellum might be involved in longer time scales of prediction as compared to motor operations. Cerebellar patients exhibit aberrant social behaviour, showing aberrant impulsivity/compulsivity. The cerebellum is a master-piece of reward mechanisms, together with the striatum, ventral tegmental area (VTA) and prefrontal cortex (PFC). Critically, studies on reward processing reinforce our view that a fundamental role of the cerebellum is to construct internal models, perform predictions on the impact of future behaviour and compare what is predicted and what actually occurs.

Simple Summary We propose a more comprehensive scheme underlying cerebellar reserve. Under pathological conditions, the internal forward model continuously updates itself to adjust the predictive control, where two reorganizing steps function cooperatively: updating predictions in the residual or affected cerebellar cortex (predictive step) and adjusting the updated predictions with the current status at the cerebellar nuclei (filtering step). Abstract Cerebellar reserve compensates for and restores functions lost through cerebellar damage. This is a fundamental property of cerebellar circuitry. Clinical studies suggest (1) the involvement of synaptic plasticity in the cerebellar cortex for functional compensation and restoration, and (2) that the integrity of the cerebellar reserve requires the survival and functioning of cerebellar nuclei. On the other hand, recent physiological studies have shown that the internal forward model, embedded within the cerebellum, controls motor accuracy in a predictive fashion, and that maintaining predictive control to achieve accurate motion ultimately promotes learning and compensatory processes. Furthermore, within the proposed framework of the Kalman filter, the current status is transformed into a predictive state in the cerebellar cortex (prediction step), whereas the predictive state and sensory feedback from the periphery are integrated into a filtered state at the cerebellar nuclei (filtering step). Based on the abovementioned clinical and physiological studies, we propose that the cerebellar reserve consists of two elementary mechanisms which are critical for cerebellar functions: the first is involved in updating predictions in the residual or affected cerebellar cortex, while the second acts by adjusting its updated forecasts with the current status in the cerebellar nuclei. Cerebellar cortical lesions would impair predictive behavior, whereas cerebellar nuclear lesions would impact on adjustments of neuronal commands. We postulate that the multiple forms of distributed plasticity at the cerebellar cortex and cerebellar nuclei are the neuronal events which allow the cerebellar reserve to operate in vivo. This cortico-deep cerebellar nuclei loop model attributes two complementary functions as the underpinnings behind cerebellar reserve.

Lesions in the Guillain-Mollaret (G-M) triangle frequently cause various types of tremors. Nevertheless, we know relatively little about their mechanisms. The deep cerebellar nuclei, representing a primary node of the triangle, have two distinct output paths: the primary glutamatergic excitatory path to the thalamus, the red nucleus, and other brain stem nuclei, and the secondary GABAergic inhibitory path to the inferior olive (IO). The excitatory path contributes to the cerebrocerebellar loop (the long loop), while the inhibitory path contributes to the cerebello-olivo-cerebellar loop (the short loop). We propose a novel hypothesis: each loop contributes to a pathophysiologically distinct type of tremors. A lesion in the cerebrocerebellar loop causes an irregular tremor. A lesion in this loop affects the cerebellar forward model. It deteriorates its accuracy of prediction and compensation of the sensory feedback delay, resulting in irregular instability of voluntary motor control. Therefore, this type of tremors, such as intention tremor or kinetic tremor, is usually associated with other symptoms of cerebellar ataxia, such as dysmetria. We call this type of tremor forward-model-related tremor. The second type of regular tremor appears to originate from the synchronized oscillation of IO cells due, at least in animal models, to reduced degrees of freedom in IO activities. The regular burst activity of IO cells is precisely transmitted along the olivo-cerebello-cerebral path to the motor cortex before inducing bursts of activities of agonist and antagonist muscles. We call this type of tremor IO-oscillation-related tremor. Although these types of regular tremors, such as essential tremor or rest tremor, do not necessarily accompany ataxia, the aberrant IO activities (i.e., aberrant complex spike, CS, activities) may induce moderate maladaptation of cerebellar forward models by reducing degrees of freedom in fundamental mechanisms of plasticity such as long-term depression (LTD) and long-term potentiation (LTP) of the cerebellar circuitry. Our hypothesis explains how lesions in or around the G-M triangle result in mixtures of two types of tremors, resulting in a complex phenotypic presentation.

This review surveys the physiological, behavioral, and computational evidence converging to the view of the cerebrocerebellum as a locus for predictive computation. State prediction is a neural mechanism that enables rapid and stable control of the body, even when sensory feedback from the periphery has a temporal delay. This predictive computation is known as an internal forward model and hypothesized to locate in the cerebellar circuit. We revisit the classical work of Marr, Albus, and Ito and argue that the cerebellum operates as a regressor of continuous dynamics rather than a classifier conventionally assumed in the Marr-Albus perceptron model. Specifically, the cerebellar output at one moment is predictive of the cerebellar input in the future, supporting the internal-forward-model hypothesis of the cerebellum. Finally, we speculate on the computation within the cerebellum and in the cerebrocerebellar loop and summarize unresolved questions for future studies.

This chapter brings enigmatic connectivity of the cerebellar dentate nucleus (DN) and the cerebellar forward model hypothesis together to demonstrate the cerebrocerebellum as loci of Kalman filters. We start with a brief history of the cerebellar internal model hypothesis. Next we present two lines of new evidence for the forward model hypothesis. First, we show physiological evidence that the cerebellar outputs from DN are predictive for the inputs to the cerebrocerebellum. Second, we introduce an enigmatic MF collateral to DN and demonstrate it is an essential key to the Kalman filter model. We further discuss how the Kalman filter model for the motor cerebrocerebellum could be generalized to non-motor parts as a unifying principle for the diverse functions of the cerebrocerebellum. We conclude that the Kalman filter model also explains how parallel modules in the cerebrocerebellar communication loops are coordinated in a cascadic manner, providing a partial explanation for unity of mind.

Lesions in the Guillain–Mollaret (G–M) triangle frequently cause various types of tremors or tremor-like movements. Nevertheless, we know relatively little about their generation mechanisms. The deep cerebellar nuclei (DCN), which is a primary node of the triangle, has two main output paths: the primary excitatory path to the thalamus, the red nucleus (RN), and other brain stem nuclei, and the secondary inhibitory path to the inferior olive (IO). The inhibitory path contributes to the dentato-olivo-cerebellar loop (the short loop), while the excitatory path contributes to the cerebrocerebellar loop (the long loop). We propose a novel hypothesis: each loop contributes to physiologically distinct type of tremors or tremor-like movements. One type of irregular tremor-like movement is caused by a lesion in the cerebrocerebellar loop, which includes the primary path. A lesion in this loop affects the cerebellar forward model and deteriorates its accuracy of prediction and compensation of the feedback delay, resulting in irregular instability of voluntary motor control, i.e., cerebellar ataxia (CA). Therefore, this type of tremor, such as kinetic tremor, is usually associated with other symptoms of CA such as dysmetria. We call this type of tremor forward model-related tremor. The second type of regular tremor appears to be correlated with synchronized oscillation of IO neurons due, at least in animal models, to reduced degrees of freedom in IO activities. The regular burst activity of IO neurons is precisely transmitted along the cerebellocerebral path to the motor cortex before inducing rhythmical reciprocal activities of agonists and antagonists, i.e., tremor. We call this type of tremor IO-oscillation-related tremor. Although this type of regular tremor does not necessarily accompany ataxia, the aberrant IO activities (i.e., aberrant CS activities) may induce secondary maladaptation of cerebellar forward models through aberrant patterns of long-term depression (LTD) and/or long-term potentiation (LTP) of the cerebellar circuitry. Although our hypothesis does not cover all tremors or tremor-like movement disorders, our approach integrates the latest theories of cerebellar physiology and provides explanations how various lesions in or around the G–M triangle results in tremors or tremor-like movements. We propose that tremor results from errors in predictions carried out by the cerebellar circuitry.

The terminology of cerebellar dysmetria embraces a ubiquitous symptom in motor deficits, oculomotor symptoms, and cognitive/emotional symptoms occurring in cerebellar ataxias. Patients with episodic ataxia exhibit recurrent episodes of ataxia, including motor dysmetria. Despite the consensus that cerebellar dysmetria is a cardinal symptom, there is still no agreement on its pathophysiological mechanisms to date since its first clinical description by Babinski. We argue that impairment in the predictive computation for voluntary movements explains a range of characteristics accompanied by dysmetria. Within this framework, the cerebellum acquires and maintains an internal forward model, which predicts current and future states of the body by integrating an estimate of the previous state and a given efference copy of motor commands. Two of our recent studies experimentally support the internal-forward-model hypothesis of the cerebellar circuitry. First, the cerebellar outputs (firing rates of dentate nucleus cells) contain predictive information for the future cerebellar inputs (firing rates of mossy fibers). Second, a component of movement kinematics is predictive for target motions in control subjects. In cerebellar patients, the predictive component lags behind a target motion and is compensated with a feedback component. Furthermore, a clinical analysis has examined kinematic and electromyography (EMG) features using a task of elbow flexion goal-directed movements, which mimics the finger-to-nose test. Consistent with the hypothesis of the internal forward model, the predictive activations in the triceps muscles are impaired, and the impaired predictive activations result in hypermetria (overshoot). Dysmetria stems from deficits in the predictive computation of the internal forward model in the cerebellum. Errors in this fundamental mechanism result in undershoot (hypometria) and overshoot during voluntary motor actions. The predictive computation of the forward model affords error-based motor learning, coordination of multiple degrees of freedom, and adequate timing of muscle activities. Both the timing and synergy theory fit with the internal forward model, microzones being the elemental computational unit, and the anatomical organization of converging inputs to the Purkinje neurons providing them the unique property of a perceptron in the brain. We propose that motor dysmetria observed in attacks of ataxia occurs as a result of impaired predictive computation of the internal forward model in the cerebellum.

Experimental results should guarantee their reproducibility for the objective nature of science. Neuroimaging data, however, often contain artifactual components that are not pertinent directly to neural activations in question, thereby impeding the reproducibility of experimental results. Signal processing or data analysis methods play a crucial role in removing such artifactual components and extracting relevant neural activations. We here provide a concise overview of data analysis methods with an emphasis on functional near-infrared spectroscopy (fNIRS) and discuss their advantages and disadvantages. Then our analysis method, task-related component analysis (TRCA), that maximizes the block-by-block reproducibility of a signal in one condition is proposed. TRCA is formulated as a generalized eigenvalue problem and is extended to several useful forms including an online recursive algorithm and one that takes channel-by-channel delays into account. Finally extensive applications of TRCA to synthetic data and fNIRS data of a finger-tapping task and a working-memory task are presented. Although originally motivated for fNIRS data analysis, the concept of signal reproducibility has a broad implication and we expect that TRCA has a wide range of applications in biophysical data analysis.