Our research partly focuses on using neuroimaging techniques to map human brain regions responsive to spectral, spatial, and temporal information.  As an example, one goal of our research is to map tonotopically organized fields in human auditory cortex.  Using a combination of a phase-encoding paradigm modeled on successful retinotopic mapping in vision, a novel hybrid electrostatic sound-delivery system, and a high resolution fMRI pulse sequence implemented at 4T, we have found evidence of multiple tonotopic maps with sharp, reversing boundaries.

Unlike previous human imaging studies of frequency maps which use only a small number of tones (2-4), we model our studies on retiniotopic mapping procedures from fMRI vision studies. In retinotopic studies, a reversing checker-pattern annulus slowly expands in diameter progressively stimulating from the peri-foveal to peripheral retina, or a checker-pattern of wedges rotates around a central fixation, thus activating corresponding retinotopically organized cortical fields.  In our auditory version, we present an AM tone, eq. 1 (cf. reversing checker pattern) that slowly increases its carrier frequency monotonically and logarithmically from 100-8000Hz (cf. expanding annulus).

       [1]

Note that the phase map shows distinct regions sensitive to different frequencies with sharp boundaries between them. Note also that within the blue shades there is evidence of a shift from light to dark (high to low) in the caudomedial to rostrolateral direction, whereas in the red shades there is a shift moving in the opposite direction, wherein relatively higher frequencies are represented rostrolaterally and lower frequencies are represented caudomedially.  This pattern is more strongly evident on the right side of the image (left hemisphere).  One interesting possibility suggested by these data is that relatively high (blues) versus relatively low (reds) frequencies are represented in distinct tonontopically organized fields.

An audio sample of the stimuli used in this experiment may be downloaded from:
http://orion.oac.uci.edu/~kourosh/nsf/wave1.wav

Motion selectivity

We are also investigating whether there exist dedicated motion-selective cortical areas. Single-cell studies have identified brainstem and cortical neurons in a variety of species that respond preferentially to a specific direction of auditory source movement, and are nonresponsive to stationary sounds. This has led to the notion that there exist brain regions specialized for encoding auditory motion.  Psychophysical evidence from human observers, however, is inconclusive, with the prevailing view questioning the existence of a separate motion mechanism.  Recent human neuroimaging data has implicated the parietal lobe and planum temporale (PT) in auditory motion processing.  These studies, however, compare a single stationary position to motion that traverses a wide range of locations.  We hypothesized that motion-based cortical activation reported for the parietal lobe and PT reflects the greater range of spatial positions covered during motion, and not the activation of a unique motion system.  We predicted that when motion-based cortical activation is contrasted to activation associated with randomly positioned sound sources covering the same range of positions as motion stimuli, no differential cortical activity will be recorded.

Results from seven subjects (columns) are shown above.  Top row shows activation patterns for motion stimuli, middle panel for stationary stimuli, and bottom row for motion minus stationary condition.  These results support the prediction that when stationary stimuli are presented randomly at multiple locations, the same activation pattern is observed as that for motion stimuli.  Since the contrast condition shows no reliable activation of previously reported motion areas.

An audio sample of the stimuli used in this experiment may be downloaded from:
http://orion.oac.uci.edu/~kourosh/nsf/wave2.wav  (requires listening through headphones).