Article Figures & Data
Figures
- Figure 1.
Clustered neuronal cultures and experimental procedure. A, top, Bright field image of a clustered neuronal culture 3 mm in diameter. Dark circular objects are neuronal clusters, and straight filaments are connections. The ablated cluster is boxed in red. Bottom, Corresponding fluorescence image after damage. Healthy clusters appear gray. The ablated cluster, with all its neurons dead, appears bright. Boxed clusters are those whose spontaneous activity is represented in panels E, F. B, Sketch of the multimodal optical system for fluorescence imaging and laser microsurgery. C, Actual field of view in the experiments. Two cultures are simultaneously monitored, with one set as control and the other one as target. The latter is the same culture as in panel A, and the red arrowhead signals the ablated cluster. D, Laser microsurgery. The four snapshots illustrate the action of the laser as it progressively scans the cluster to be ablated, delivering in each step a high energy, high penetration pulse that kills the neurons and vaporizes water. The time interval between panels is 20 s. E, Spontaneous activity before damage for the five clusters highlighted in A. Activity is rich and all clusters fire together in a highly coordinated manner. The red arrowhead marks the cluster to be ablated. F, Corresponding activity after damage, with the ablated cluster completely silent. Its immediate neighbors are initially silent but recover activity after ∼10 min, although with lower firing rates and amplitudes (black arrowheads). Clusters more distant from damage maintain their activity after ablation, although with a reduced firing rate.
- Figure 2.
Network evolution during recovery. A, Network effective connectivity and raster plot of activity before damage. The effective connectivity is computed on the full, 30 min duration of the recording. Nodes and links are color coded according to their strength and weight, respectively. The darker the color, the higher the value. The yellow arrowhead marks the targeted cluster.
provides the global efficiency before damage. The bottom raster plot shows the 10 min before damage, with the yellow band highlighting the cluster to be ablated. Black dots are activations. B, Effective connectivity evolution and raster plot after damage. The effective connectivity networks were computed in ∼6 min time windows. The ablated cluster is marked in black. Clusters in blue are those that became silent just after damage but recovered afterward, with the numbers indicating their location in the bottom raster plot. G provides the global efficiency, and its relative change with respect to
is shown in brackets. In the raster plot, the ablated cluster is shown with a yellow band; the initially silent clusters are shown with a white band. One of these clusters never recovered and the band encompasses the full duration of the raster plot. Gray dots are activations in clusters that did not substantially change activity after damage. Black dots are activations in affected clusters. - Figure 3.
Variation of the global efficiency and the density of links on damage. A, top plot, Time evolution of the global efficiency G for control (green) and ablated (black) cultures, before and after damage. Bottom plot, Corresponding density of links D. The yellow panels provide representative effective networks of the experiment shown in Figure 2, are computed over ∼5-min time windows (gray horizontal bars) and are thresholded to show the 5% links with the highest weight. The ablated cluster is marked in red. The networks illustrate the important changes in the distribution of links’ weights along the recovery process. In both plots, data were averaged over 14 cultures and the shadings show SD. For clarity, only the last 10 min before damage and the first 15 min after damage are shown. B, Box plots of the distribution of G (top) and D (bottom) values for the 14 cultures at different experimental conditions, comparing controls before and after damage with ablated cultures before damage, the first 5 min after ablation (aft0) and the last 15 min of the recording and that correspond to the recovered state (aftR). For both G and D, significance (***p < 0.001, one-way ANOVA) is only observed between the condition just after damage and the rest of conditions. C, Evolution of the global efficiency for a representative individual experiment to spotlight the definitions of the global efficiency loss Λ, recovery time RT, and recovery rate Θ.
and
are, respectively, the global efficiencies before damage (dotted line for average, gray shading for SD) and just after damage. D, Distributions of Λ, RT, and Θ for all 14 experimental realizations. All box plots span from the median to the first and third quartiles, and whiskers span from the 10th to 90th percentile. - Figure 4.
Spontaneous activity in the neighborhood of damage. A, Construction of the ring of neighbors for a representative culture. The first neighborhood ring (red clusters) is constituted by all clusters whose centers fall within a distance
mm (yellow circle) from the ablated cluster (gray). The second and subsequent rings are built by identifying the clusters that are neighbors of the clusters in the previous ring according to the same distance RC. B, Box plots showing the temporal evolution of the average activity in six neighborhoods and for the 14 experimental realizations. Before damage, activity is averaged over 30 min. The indicated times correspond to the center of the analysis windows t. For damage (at t = 3 min, boxed) and subsequent recovery stages, activity is averaged in 6-min windows. Dotted black arrows highlight two experiments whose activity boosted up after damage. Average values of the distributions are shown as mean ± SD. The colored panels highlight the distributions that are significantly different according to a one-way ANOVA (*p < 0.05, **p < 0.01). - Figure 5.
Network communication during recovery. A, Interaction probability P among all pairs of neighborhoods. The brighter the color, the higher the formation of effective links between and within neighborhoods. Data were computed for each culture and then averaged over the 14 studied cultures. The left panel shows the neighbors’ interaction before damage, with data averaged over 30 min. The three central panels show the action of damage and subsequent recovery, with each panel corresponding to ∼6-min window intervals for analysis. The last panel shows the stationary recovery, with data averaged over a broader window of ∼12 min. B, Corresponding representation of the flow of links
(before damage) and the percentage variation of flow of links F* (rest of panels, calculated with respect to the predamage scenario). The blue curved arrow and the numbers indicate the distance from damage in terms of neighborhoods. Damage locus is symbolized as a red band. Before damage, purple arrows depict the communication flow between neighborhoods. After damage, the arrows depict the level of formation of new effective links between neighborhoods with respect to predamage. Arrows’ thickness and color intensity are coded according to the values of F or F*. Nodes’ color is coded according to the relative strength (weighted sum of incoming and outgoing links) of the neighborhood.
In this issue
