COMPUTATIONAL THEORY FOR MOVEMENT PATTERN-RECOGNITION BASED OPTIMAL MOVEMENT PATTERN GENERATION

We have previously proposed an optimal trajectory and control theory f or continuous movements, such as reaching or cursive handwriting. Acco rding to Marr's three-level description of brain function, our theory can be summarized as follows: (1) The computational theory is the mini mum torque-change model; (2) the intermediate representation of a patt ern is given as a set of via-points extracted from an example pattern; and (3) algorithm and hardware are provided by FIRM, a neural network that can generate and control minimum torque-change trajectories. In this paper, we propose a computational theory for movement pattern rec ognition that is based on our theory for optimal movement pattern gene ration. The three levels of the description of brain function in the r ecognition theory are tightly coupled with those for pattern generatio n. In recognition, the generation process and the recognition process are actually two flows of information in opposite directions within a single functional unit. In our theory, if the input movement trajector y data are identical to the optimal movement pattern reconstructed fro m an intermediate representation of some symbol, the input data are re cognized as that symbol. If an error exists between the movement traje ctory data and the generated trajectory, the putative symbol is correc ted, and the generation is repeated. In particular, we present concret e computational procedures for the recognition of connected cursive ha ndwritten characters, as well as for the estimation of phonemic timing in natural speech. Our most important contribution is to demonstrate the computational realizability for the 'motor theory of movement patt ern perception': the movement-pattern recognition process can be reali zed by actively recruiting the movement-pattern formation process. The way in which the formation process is utilized in pattern recognition in our theory suggests a duality between movement pattern formation a nd movement pattern perception.