Figure 5.

Organization of evoked network activity in the AL. (A) Illustration of the trajectories representing PN population activity over time in response to cis-3-hexen-1-ol (green), linalool (blue), and citral (red). For each odor, 47 PN spike sequences were divided into consecutive 50-ms time bins and the number of spikes in each bin was counted. We viewed the 47-dimensional vector as odor representation by the PN population. To visualize the 47-dimensional vector, we applied principal component analysis (PCA) to the vector and used the first three PCs for the graph. The variance captured by the first three PCs was 54.0% of the total. Each representation evolved in an odor-specific manner. (B) Euclidean distance between odor representations as a function of time. We calculated the Euclidean distance between the 47-dimensional vectors used in (A). Each trace showed similar profiles. The distances increased by about 100 ms after stimulus onset, peaking at 250–300 ms after stimulus onset (150–200 ms after response onset). At 550–750 ms, possibly in sustained stimulus presentation, the distances began to stabilize. From about 1000 ms after stimulus onset, the distances remained stable and at higher levels than before stimulation. (C) Distribution of factor loadings for PCs 1–3. Glomeruli are ordered roughly from anterior to posterior (right to left in the graph). We applied PCA to the 32-dimensional vector as in (A). The firing rates of PNs innervating the same glomerulus were averaged. (D) Response similarity as a function of anatomical distance between glomeruli. We used Pearson correlation coefficients of the responses of pairs of PNs to represent response similarity. Firing rates from the stimulus onset to 1 s after the stimulus onset (20 time bins) were used (the same value used in Figures 2D, 2E, and 3B). Pairs of PNs innervating different glomeruli show no strong correlation (r = –0.0556, n = 496). (E) Anterior and posterior views of the color map of the olfactory response profile. First, we calculated correlation coefficients for all pairs of 32 glomeruli (32 × 32 matrix). To visualize the 32-dimensional vectors, PCA was used and vectors were embedded in the color space defined by the first three PCs (magenta, blue, and yellow for PCs 1–3, respectively). Glomeruli were labeled by the color defined in the PC space. PCs 1–3 account for 79.5% of the total variance. Similarly colored glomeruli have similar response profiles; differently colored glomeruli have different response profiles.