[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: defining a 'fixation point' in audition

Hi Francesco,

In visual attention studies, the need for a fixation point arises from the nature of vision:Â
-- First, the eye can move.
-- Second, not all parts of the eye are equally sensitive.
-- Third, the visual system can focus its attention on a position in space.Â

If you didn't use a fixation point, then given the unpredictable movements of the eye, and assuming that the research involved a briefly flashed image, you wouldn't know which part of the eye was going to get this image, or where attention would be focused at the moment of onset of the image.

The analog in hearing involves similar considerations:
-- Although the ears of humans can't move, the movement of the head can change their positions. (Not important wen you use headphones)
--Â The ears are more sensitive to sounds coming from some directions than others.
--Â Auditory attention can be focused on a spatial position.

Since in humans, the ears don't move, immobilizing the head will guarantee the angle of arrival of the signal. It can also be used with headphones where it will guarantee that perceived location in the headphone space and perceived location in the space of the surrounding room are strictly correlated, so that you don't have to worry about which type of location the subject was attending to. If you want to _control_ spatial attention, you could play a sound (perhaps a click or a soft noise burst) from the location you wanted to have as the focus of attention before each trial.

However, I assume that the auditory system can also focus its attention on a pitch, a frequency range, or a time interval. You haven't described the nature of your task. If you wanted to focus auditory attention on some value of one of these features of the signal before each trial, you could do so by playing a sound having that value before the trial (e.g., a pure tone of a certain pitch and location). If you just want to give a subject a "get ready" signal, I have found that a short soft burst of white noise at a fixed time interval before the target signal works well. You don't want it to be too loud or too close in time to the target signal if you want to avoid partial forward masking of the target by the warning signal and avoid triggering an "alarm" response in the listener with an overly loud sound. A one-second interval is good. If you wanted to employ rhythmic information to ready the subject for the stimulus, you could use a sequence of two ready signals (A and B) preceding the onset of the target (T),. In the sequence A-B-T, the A-B onset-to-onset interval would be the same as the B-T onset-to-onset interval.

Albert S. Bregman, Emeritus Professor
Psychology Department, McGill University
1205 Doctor Penfield Avenue
Montreal, QC, Canada H3A 1B1.
Office:Â Phone: (514) three-nine-eight-6103,
      Fax: (514) three-nine-eight-4896

On Thu, May 3, 2012 at 3:40 PM, ftordini@xxxxxxxxx <ftordini@xxxxxxxxx> wrote:
Dear list members,
I am in the process of redesigning a test dealing with attention and auditory
saliency but then I have a doubt about how to define, and possibly control, the
'fixation point' (in the case of sound stimuli).
I am using headphones, so head-tracking is not an option, I believe.

I also have thought that, giving credit to a supramodal model of attention, I
could provide a visual fixation point (a cross on the screen) but this would be
a fallback solution since I would like to keep it unimodal at this stage (and I
am addressing stimulus-driven attention).

This may sound as a very naive question to most of you, but I am turning it to
the community since I always got precious comments and feedback in the past.

Best regards

Francesco Tordini   Ââ skype francesco.tordini
PhD Student       â http://www.cim.mcgill.ca/sre/personnel/

Centre for Intelligent Machines and
Department of Electrical and Computer Engineering
McGill University
3480 University Street, Room 426
Montreal, QC, H3A 0E9, Canada Â| T: (514) 398 8201