| Abstract: | Pattern recognition with a prosthesis substituting vision by audition was investigated. During 15 1‐hour sessions, nine blindfolded sighted subjects were trained to recognise 2D patterns by trial and error. In addition to a global assessment, recognition of pattern element nature (vertical bars, horizontal bars…), element size and element spatial arrangement were independently assessed for each pattern. Influence of experimental parameters (complexity level of patterns, exploration number of a pattern) on recognition was studied. Performances improved over sessions. As a rule, patterns element nature was less well recognised than element size and spatial arrangement. Experimental parameters influenced pattern recognition performance. Results are discussed in relation with auditory and visual perception as well as in the perspective to implement a learning protocol for future users of prosthesis. Copyright © 2006 John Wiley & Sons, Ltd. |
