====== Manuscript in preparation ========= Evidence for effector independent and dependent components and their differential time course of acquisition during motor sequence learning Running head: Visual and motor representations for sequence learning Authors: Raju S. Bapi and Kenji Doya Computational Neurobiology Group, Kawato Dynamic Brain Project, ERATO, JST, 2-2 Hikaridai, Seika, Soraku, Kyoto, Japan 619-0288 Fax: +81 774 95 3001 Tel: +81 774 95 1211/1210 Email: {rajubapi,doya}@erato.atr.co.jp Abstract We tested transfer effects in a motor sequence learning paradigm in order to investigate the representation of motor sequence. It is hypothesized that sequence representation is mediated by two components -- effector independent and dependent. Further, it is postulated that the effector independent component is in visual/spatial coordinates and the effector dependent component is in motor coordinates and that they have different time courses of acquisition during learning. Six subjects were tested in a modified 2x10 task (Hikosaka et al. 1996a; Sakai et al. 1998). Subjects learned to press two keys (called a set) successively on a keypad in response to two lighted squares on a 3x3 display. Ten such sets, called a hyperset, were used. Training was given in the NORMAL condition and sequence recall was assessed in the early and late stages in three conditions -- NORMAL, VISUAL, and MOTOR. In VISUAL condition access of the arm to the keypad was rotated 90 degrees while the keypad-display mapping kept identical to NORMAL. In MOTOR condition the keypad-display mapping was also rotated 90 degrees, resulting in an identical finger-display mapping as in NORMAL. Reaction time (RT) and errors during recall were recorded. Errors did not significantly differ based on position, showing that difference in effector movements did not affect recall of sequence knowledge. This result supports an effector independent sequence representation, possibly in visual/spatial coordinates. With training, RTs in NORMAL and MOTOR conditions became similar and significantly lower than in VISUAL condition. Using RTs for single key pressing in the three conditions as baseline indices, it was observed that RTs improved significantly more in NORMAL and MOTOR conditions by the late stage of training. In contrast, RTs in VISUAL condition improved less, supporting the hypothesis that MOTOR condition benefits more than VISUAL because it uses identical effector movements to the NORMAL condition. This result supports an effector dependent sequence representation in motor coordinates that develops relatively slowly. The difference in the time course of learning of these two representations may account for the differential involvement of brain areas in early and late learning phases found in lesion and imaging studies. Key words: Sequence representation, Effector dependence, Effector independence, Procedural learning, Transfer of sequence knowledge, Time course of sequence learning