Speechreading in the akinetopsic patient, L.M.

Patient L.M. has a well-documented, long-standing and profound deficit in the perception of visual movement, following bilateral lesions of area V5 (visual movement cortex). Speechreading was explored in this patient in order to clarify the extent to which the extraction of dynamic information from facial actions is necessary for speechreading. Since L.M. is able to identify biological motion from point-light displays of whole-body forms and has some limited visual motion capabilities, we expected that some speechreading effaces in action would be possible in this patient. LM.'s reading of natural speech was severely impaired, despite unimpaired ability to recognize speech-patterns from face photographs and reasonable identification of monosyllables produced in isolation. She was unable to track multisyllabic utterances reliably and was insensitive to vision when incongruent audiovisual speech syllables were shown. Point-light displays of speech were as poorly read as whole face displays. Rate of presentation was critical to her performance. With speech, as with other visual events, including tracking the direction of gaze and of hand-movement sequences, she could report actions that unfolded slowly (similar to one event per 2 s). In line with this, she was poor at reporting whether seen speech rate was normal, fast (double-speed) or slow (half-speed). LM.'s debility is the converse of that reported for a patient with lesions primarily to V4 (H.J.A.), who is unable to speechread photographs of faces but can speechread moving faces. The visual analysis of both form and motion is required for speechreading; the neural systems that support these analyses are discussed.