Fig. 6.
Grouping of sweep trajectory decoding across SNRs for all sites (n = 321). Three SNR response classes were defined. A: monotonic: sites whose decoding performance was maximal in the absence of noise and greater than that obtained in the presence of noise at any of the tested SNR conditions (n = 104). B: nonmonotonic: sites whose decoding performance is maximal in the presence of noise at one of the tested SNR conditions and significantly greater than that obtained in the absence of noise (n = 101). C: equivalent: Sites whose decoding performance in the absence of noise is not statistically different from performance in the presence of noise for at least one tested SNR condition (n = 116). Waterfall plots show performance as function of SNR for all sites in each response class. Untested SNR conditions of −25 dB were set to zero for graphical purposes. Other untested SNR conditions were linearly interpolated on the basis of existing values. Coloring varies from the lowest (blue) to highest (red) values on each panel. D: distribution of the different response groups for the 5 different SNR conditions. Blue slices represent performance at the indicated SNR less than that for the no-noise (SNR+∞) condition. Red slices represent performance better than that for SNR+∞. Green slices represent performance indistinguishable from SNR+∞. Gray slices represent sites with no significant decoding performance at any tested SNR. Because not all SNR conditions were tested at every site, the data from −20 to 0 dB were obtained at 307, 273, 307, 295, and 316 sites respectively.