FIGURE 5.

Synchronous bursting events dynamics in a PKC^- network at a 24 DIV (NW₃). (A) SBEs were initiated in a much narrower region than in PKC^N networks and propagated across the network in a much more homogeneous fashion (3 × 3 median filter smoothing). White crosses mark the first ten recruited electrodes in a sample SBE and the black spot their means of the x and y coordinates defined as onset location. (B) Onset locations were located predominantly along the boundary and formed distinct BIZs (N = 486 SBE, t = 103 min; 4.7 SBE/min). Contour lines indicate the frequency with which individual electrodes were among onset electrodes (smoothed by 3 × 3 median filtering). (C) BIZs were identified by spatial centroid clustering of onset locations (cut-off at 1 mm distance between onset locations). (D) Burst initiation was dominated by even fewer BIZs than in PKC^N networks. The nine most frequent BIZ are color coded. (E) Mapping individual SBE onset locations shown in B to their respective BIZ (color code as in D) indicates that dominating BIZ lay close to but not at the network boundary. (F) BIZs reflected the centers of burst onset regions, which were mostly compact but not confined to extremely localized positions. The maps show the probability by which electrodes were among the first ten onset electrodes of bursts starting within the dominating BIZs. (G) As in PKC^N networks, average relative activity levels at BIZ electrodes (ratio of the mean AFR at BIZ electrodes and of all other electrodes with spike activity) were slightly above network average in the dominating BIZs and always lower than the 25% of highest AFRs. (H) Map of relative activity levels (ratio between the AFR at individual electrodes and network AFR during SBEs). BIZs appeared mostly located between hot and cold spots. Note that the large central region with high relative activity levels never initiated SBEs. (I) Median burst strength at BIZ electrodes when driving SBEs (active) or recruited during SBEs initiated by other BIZs (passive). Activity in the major BIZs was not significantly higher when they initiated SBEs. (J) Average propagation patterns elicited by the first nine BIZs revealed a homogeneous propagation of activity from different BIZ positions. (K) Similarity between propagation patterns was determined as the correlation of electrode recruitment ranks during SBEs. Sorting according to BIZ assignment revealed a very high correlation between propagation patterns originating at the same BIZs. (L) As in PKC^N networks, the correlation between propagation patterns dropped as function of distance between BIZs but yielded highly anti-correlated patterns for BIZ located at opposed sites of the network. (M) The distribution of SBE onset locations for the same network recorded one week later was highly similar (30 DIV, N = 627 SBE, t = 140 min; 4.5 SBE/min). (N) The overall distribution of BIZs was mostly preserved during development. Comparing letter sequence (reflecting decreasing frequency) to the numbers in H shows, however, that their influence had changed, focussing SBE initiation even more on one dominating BIZ. (O) Relative activity levels across the MEA displayed highly similar patterns at 24 and 30 DIV.