Re: By any other name... (Israel Nelken )


Subject: Re: By any other name...
From:    Israel Nelken  <israel@xxxxxxxx>
Date:    Fri, 23 Mar 2007 09:22:24 +0200
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

To follow Daniel Pressnitzer's message: in bad SNR the presence of the tone can be difficult to establish, because, well, of the presence of the noise. Any objective method for deciding whether a tone was there during the noise segment will run into the absolute limits imposed by 'ideal observer analysis', based on whatever measure you think is relevant (spike counts, intervals, absolute spike times...) in the appropriate channels. Thus, at the rather bad SNR in which the continuity illusion is usually run, an ideal observer would probably be unable to discriminate between noise+tone and noise alone during the noise burst with any reliability. It's only the experimenter who may know with certainty whether the tone was there, not the listener. Thus, I would strongly support Daniel's idea that the role of the noise is to make the situation ambiguous, rather than to supply the information about the presence of the tone. Under these circumstances, heuristics (such as the 'old+new') probably become important in the higher-level interpretation of the scene. For what it is worth, here's my own self-promotion: we did a study of the neural responses to low-level tones in fluctuating maskers (Las et al. J. Neurosci. 2005). The parameters were set for studying comodulation masking release rather than the continuity illusion, but with this caveat, we found that in cortical neurons responded to the noise+tone stimulus similarly to their responses to tone alone (strongly suppressing the firing rate fluctuations that correlated with the noise envelope). At the same SNRs, this was not the case in inferior colliculus - changes in firing patterns were enough to signal the presence of the tone, but the responses were still more similar to the noise alone responses. Moving to unsupported speculations, we concluded that the representation of the tone as a separate 'object' was substantially enhanced in cortex relative to the inferior colliculus. Maybe this is a neural signature of the perceptual continuity of the tone within the noise. Eli -- ================================================================== Israel Nelken Dept. of Neurobiology The Alexander Silberman Institute of Life Sciences Edmond Safra Campus, Givat Ram | Tel: Int-972-2-6584229 Hebrew University | Fax: Int-972-2-6586077 Jerusalem 91904, ISRAEL | Email: israel@xxxxxxxx | *NOTE CHANGE IN EMAIL ADDRESS!* ==================================================================


This message came from the mail archive
http://www.auditory.org/postings/2007/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University