The reconstruction of coding scheme through errors distributions Lyakhovetskii V.A., Karpinskaya V.Ju*, Bobrova E.V. Pavlov Institute of Physiology of Russian Academy of Science * Saint-Petersburg State University

Introduction At early stage of memorization the coding schemes are tightly connected with the structure of external space (Веккер, 2000). If we use the stimuli permitting to estimate quantitatively the distance between correct and wrong answers then we may calculate some errors distributions. Main hypotheses. Depending on spatial relations code the distortion of its elements leads to different errors distributions. The simulation enables to correlate different spatial errors distribution to the coding schemes (Ляховецкий и др., 2006). Stimuli types visually presented sequences of chess pieces or of targets for hand movements (Ляховецкий и др., 2006; Bobrova et al., 2011) kinaestheticaly presented hand movements sequences (Ляховецкий, Боброва, 2009) visually presented single stimulus for hand movement

Experiment 1. Spatial transformations of chess positions Variants of piece transitions: (A) move according to the rule, (B) free-form. Experimental design The recognition stage of experiment

Results 1. Spatial transformations of chess positions The errors distributions for two types of stimuli transitions (free-form and move variants) are different. The exponential approximation is significant only for the move according to the rule variant (moving bishop). The model developed to simulate the results uses two distinct coding schemes: 1) vector (relative) coding: for the information about stimuli position which are stored in relation to the preceding one (move according to the rule variant) 2) positional (absolute) coding: all stimulus are stored independently from each other (free-form variant)

Experiment 2. Hand movements sequences Memorization Kinaesthetically presented sequence: The hand of the blindfolded volunteer was moved by experimenter from center through 6 different positions on a sheet of paper A4 located under touch screen. OR Visually presented sequence: The volunteer saw six consecutive numbers in the cells of a 6x4 table presented on the display. Reproduction The volunteer had to immediately reproduce the sequence of positions by the right or the left hand

Results 2. Hand movements sequences The errors distributions for visually presented stimuli are log-normal. The errors distributions for kinaesthetically presented stimuli are exponential – the exponential decay is more accentuated for right hand. The model predictions: visually presented stimuli: positional coding (Bobrova et al., 2011) kinaestheticaly presented stimuli: right hand – vector coding left hand - vector coding + positional coding (Ляховецкий, Боброва, 2009)

Experiment 3. Single stimulus for hands movement Experimental design Visual stimuli 2 groups of right-handed volunteers (25 volunteers in each). One group begins the task with the left hand, another group – with the right one. 5 different shafts length: 5 cm, 6.6 cm, 8.3 cm, 11.6 cm Experimental results (example)

Single stimulus for hands movement Typical errors distributions - the exponential decay is more accentuated for right hand.

Single stimulus for hands movement For right hand of R group observers reproducing the upper shafts the errors distributions are significantly more exponential than for the left hand. R group, right hand: vector coding first, then vector coding + positional coding R group, left hand; L group vector coding + positional coding

Conclusions the errors distributions reflect essential features of coding schemes the distinctions of errors distributions may be detected in investigations of memorization of both sequence and single stimulus data support the hypothesis about using relative or absolute coding schemes (vector or positional) for memorization of sensory information and control of the movements of the right and the left hands