Functional topography of cat primary auditory cortex: responses to frequency-modulated sweeps

Abstract

The spatial distribution of neuronal responses to frequency-modulated (FM) sweeps was mapped with microelectrodes in the primary auditory cortex (AI) of barbiturate-anesthetized cats. Increasing and decreasing FM sweeps (upwardand downward-directed FM sweeps, respectively) covering a range of 0.25–64.0 kHz were presented at three different rates of frequency change over time (i.e., sweep speed). Using multiunit recordings, the high-frequency domain (between 3.2 and 26.3 kHz) of AI was mapped over most of its dorsoventral extent (as determined by the distribution of the excitatory bandwidth, Q_10dB) for all six cases studied. The spatial distributions of the preferred sweep speed and the preferred sweep direction were determined for each case. Neuronal responses for frequency sweeps of different speeds appeared to be systematically distributed along the dorsoventral axis of AI. In the dorsal region, cortical cells typically responded best to fast and/or medium FM sweeps, followed more ventrally by cells that responded best to medium — then slow-, then medium-speed FM sweeps. In the more ventral aspect of AI (which in some cases may also have included cells located in the dorsal region of the second auditory field, AII), neurons generally preferred fast FM sweeps. However, a comparison of maps from different animals showed that there was more variability in the distribution of preferred speed responses in the ventral region of the cortex. The directional preference of units for FM sweeps was determined for the sweep speed producing the strongest response. Direction selectivity appeared to be nonrandomly distributed along the dorsoventral axis of AI. In general, units that responded best to upward-directed FM sweeps were located in the more dorsal and ventral aspects of AI while units that responded best to downward-directed FM sweeps were usually located in the mid-region of AI. Direction selectivity was also determined for multiunit responses at each of the three FM sweep speeds. In general, there was a relatively close agreement between the spatial distributions of direction selectivity determined for the strongest response with those calculated for the fast and medium speeds. The spatial distribution of direction selectivity determined for slow FM sweeps deviated somewhat from that determined for the strongest response. Near the dorsoventral center of the mapped areas, the distribution of units that responded best to downward sweeps tended to overlay the distribution of units that responded best to slow speeds, suggesting some spatial covariance of the two parameters. However, when the analysis was extended over the entire region of cortex examined in this study, the point-by-point correlation between preferred speed and direction selectivity was not statistically significant. In addition, when neural responses obtained from the dorsal and ventral subregions were analyzed separately, no significant correlation was observed between these two response parameters. This suggests that, for a given cortical location, the response properties of direction selectivity and preferred speed are derived from distinct neural processing mechanisms. Significant observations were also made between preferred FM sweep speed and excitatory bandwidth (i.e., Q_10dB and Q_40dB) such that units that responded best to slower FM speeds also seemed to have higher Q_10dB and Q_40dB (i.e., were narrowly tuned) and vice versa. In addition, units that responded well to a broadband transient stimlus in general preferred faster FM sweeps and vice versa. Although these correlations were significant across the entire dorsoventral extent of AI investigated in this study, they were stronger for responses in the dorsal subregion of AI. For direction selectivity, statistically significant correlations with these response parameters were observed more often in the dorsal than the ventral regions of AI. The apparent spatial segregation of neuronal responses to different FM sweep speeds and sweep directions distributed along the isofrequency domain of AI suggests that the global aspects of cortical function are compatible with psychophysically derived notions of parallel streams of processing for different aspects of FM signals.