Nine infants were tested, at the age of onset of reaching, seated on their parent’s lap and reaching for a small plastic toy. Kinematic analysis revealed that infants largely used shoulder and torso rotation to move their hands to the toy. Many changes in hand direction were observed during reaching, with later direction changes correcting for earlier directional errors. Approximately half of the infants started many reaches by bringing their hands backward or upward to a starting location that was similar across reaches. Individual infants often achieved highly similar peak speeds across their reaches. These results support the hypothesis that infants reduce the complexity of movement by using a limited number of degrees-of-freedom, which could simplify and accelerate the learning process. The proximodistal direction of maturation of the neural and muscular systems appears to restrict arm and hand movement in a way that simplifies learning to reach.

This paper describes the current state of our exploration of how motor program concepts may be related to neural mechanisms. We have proposed a model of sensorimotor networks with architectures inspired by the anatomy and physiology of the cerebellum and its interconnections with the red nucleus and the motor cortex. We proposed the concept of rubrocerebellar and corticocerebellar information processing modules that function as adjustable pattern generators (APGs) capable of the storage, recall, and execution of motor programs. The APG array model described in this paper extends the single APG model of Houk et al. (1990) to an array of APGs whose collective activity controls movement of a simple two degree-of-freedom simulated limb. Our objective was to examine the APG array theory in a simple computational framework with a plausible relationship to anatomy and physiology. Results of simulation experiments show that the APG array model is capable of learning how to control movement of the simulated limb by adjusting distributed motor programs. Although the model is based on many simplifying assumptions, and the simulated motor control task is much simpler than an actual reaching task, these results suggest that the APG array model may provide a useful step toward a more comprehensive understanding of how neural mechanisms may generate motor programs.

The correlation of the activity of neurons in the interposed and dentate nuclei of the cerebellum with conditioned movements of the nictitating membrane was investigated using linear systems analysis. The activity of single deep cerebellar nuclear cells was assumed to be the input to a linear system that produced nictitating membrane movement. Data were initially analyzed with a causal model to assess the degree to which past neural activity predicted the conditioned response. 55 of 165 cells had correlation coefficients of 0.50 or greater between the model's moment-to-moment output and the actual output, with two interpositus cells having correlation coefficients of greater than 0.90. Double-sided impulse responses indicated that afference from the face and efference copy probably affect deep cerebellar neural activity. Nonlinearities were also found in the relationship between neuronal activity and conditioned movement. It was concluded that cerebellar deep nuclear firing is highly correlated with future nictitating membrane movements but that the firing-movement relationship contains noncausal and nonlinear components.

The activity of neurons in the interposed and dentate nuclei of the cerebellum was investigated during differential classical conditioning of the rabbit eye blink nictitating membrane response. Forty-seven percent of the 165 cells in the study responded to the orbital stimulation used as the unconditioned stimulus (US). The latency distribution of US-elicited responses was bimodal with peaks at 7 and 19 ms. Twenty-one percent of the cells responded with short latencies to the tones used as conditioned stimuli (CSs). These cells typically responded to both the reinforced and nonreinforced CSs. Forty-one percent of the cells responded on conditioned response (CR) trials but not on trials without CRs. The average lead of the neural response to the CR was 71.4 ms. Cells that responded on CR trials were more likely to respond to the CSs, or to the CSs and the US, than cells that did not respond on CR trials. For about half of the cells that responded on CR trials the latency of response followed trial-by-trial variations of CR latency. For the remainder, the response was time-locked to CS-onset. Cells whose responses paralleled the CR may be involved in the initiation or modulation of the CR, while those whose responses were time-locked to the CS may be involved in sensory processing underlying the initiation of the movement. The pathways that may underlie the US- and CS-elicited responses are also discussed.

Conditioned responses often reflect knowledge about the timing of a US. This knowledge is manifested in the dependance of response topography on the CS-US interval employed in training. A neural network model and set of learning rules capable of simulating temporally adaptive features of conditioned responses is reviewed, and simulations are presented. In addition, we present a neural network implementation of the model which is designed to reconcile empirical studies of long-term synaptic depression in the cerebellum with neurobiological evidence from studies of the classically conditioned nictitating membrane response of the rabbit.

In 1980 our colleague John E. Desmond (Desmond et al., 1981) discovered that a small unilateral lesion of the dorsolateral pons (DLP) could completely and irretrievably eliminate a previously acquired conditioned response (CR). As in all our behavioural studies, the CR was the classically conditioned nictitating membrane (NM) response of the rabbit. The NM response is a passive consequence of defensive or protective retraction of the eyeball. In a typical experiment, the animal is given 100 pairings/day of a tone conditioned stimulus (CS) with mild electrostimulation of one eye as an unconditioned stimulus (US). The onsets of the two stimuli are separated by a few hundred msec, and these pairings or trials are separated by 15-30 sec. Conditioned responses are extensions of the NM occurring within the CS-US interval. A CR typically consists of a graded movement of the membrane in a naso-temporal direction with full extension coinciding with US onset. CRs typically begin to emerge within a hundred or so trials. Desmond’s discovery was accidental because he was attempting to lesion the locus coeruleus, a structure lying just medial to the lesion that eliminated CRs. Earlier lesion and ablation studies had shown that brain regions rostral to the midbrain are not essential for NM conditioning (Moore, 1979; Moore et al., 1980; see also Mauk and Thompson, 1987).