Subject: Streams From: Al Bregman <BREGMAN(at)PSYCH.MCGILL.CA> Date: Fri, 18 May 2001 13:32:33 -0400Dear List, I would like to comment on some characterizations of stream segregation offered recently by Bruno Repp and to present my own, by way of contrast. According to Bruno, "I believe the difference is that groups are sequential whereas streams are interleaved: xxxxx ooooo (groups) xoxoxoxoxox (streams)" Bruno's definitions are different from the ones that I use. For me, a stream is the brain's representation of the sound emitted by a single sound source over time. It doesn't matter whether the sound is intermittent or continuous, interleaved with another sound or not. A very long steady unbroken tone against a background of silence is a stream. There are, of course, variations in the structures of streams (e.g., steady vs. intermittent, temporally random vs. rhythmic, etc.) and in the presence of other sounds. I don't believe that the choice of the definition that I use is arbitrary. If all these types of streams have common factors that influence them and common consequences of their formation -- as I believe they do -- this justifies treating them as variants of the same type of mental representation. Bruno also wrote: "Stream segregation, I believe, is a phenomenon revealed by explicit judgments of the auditory scene that are made either directly (by stating how many streams are heard) or indirectly (by judging timbre, timing, or the like). In other words, streams are conscious parsings of the auditory input, or parsings that have consequences in conscious perception." There is an issue about whether streams are necessarily conscious. Obviously when a listener is asked to made a judgement about a stream or about a sound within a stream, the brain has to create a conscious representation of it. However I believe that the same grouping processes that form streams also operate on sound that is not being attended to by the listener. Studies using the mismatch negativity paradigm with evoked EEG potentials, carried out by Elyse Sussman at the Albert Einstein School of Medicine, seem to show that this is so. Attention seems to strengthen the stream segregation but not to be essential for obtaining it. An old study I did with Alex Rudnicky appeared to show that the perceptual organization of sounds that listeners were trying to exclude from attention made it easier or harder for them to do so. Therefore they were organized without being at the focus of attention. Bob Carlyon has evidence that seems to show that attention is a prerequisite for streaming; so the question is not resolved. However there is a strong possibility that grouping is pre-attentive. Still, it is likely that we do not perceive any conscious streams until this presumed preattentive grouping gets together with top-down processes to produce the conscious representation. It is possible -- even likely -- that in some cases, top-down processes can over-ride the groupings that result from bottom-up grouping. Al ------------------------------------------------- Albert S. Bregman, Emeritus Professor Dept of Psychology, McGill University 1205 Docteur Penfield Avenue Montreal, QC, Canada H3A 1B1 Office: Phone: +1 (514) 398-6103 Fax: +1 (514) 398-4896 Home: Phone & Fax: +1 (514) 484-2592 Email: bregman(at)psych.mcgill.ca -------------------------------------------------