Experimental timeline and methods.

(A) Timeline of experimental procedures, imaging times, and behavioral tests. (B) Glass cranial windows and headplates were implanted 2 weeks prior to beginning habituation protocols in Thy1-GCaMP6S mice. (C) Mice were habituated on the floating homecage, string pull and tapered beam for a period of 2 weeks prior to baseline measurements for behavior. Widefield Ca2+ response maps were used to determine locations for the imaging site located between the cFL and cHL sensory maps at baseline (“peri-infarct” imaging region) and imaging site lateral to the cHL at baseline (“distal” imaging region). For each Ca2+ imaging session, both imaging regions were imaged for a period of 15 minutes while simultaneously tracking animal movement within the mobile homecage. After baseline Ca2+ imaging, the somatosensory cFL map was targeted with photothrombosis. Behavioral testing on the string pull task and tapered beam task were performed weekly 1 day prior to the Ca2+ imaging session. After the final cellular Ca2+ imaging session at the 8 week time, stimulus- evoked widefield Ca2+ imaging was once again performed to determine the cFL and cHL somatosensory maps. (D) Representative widefield and two photon calcium imaging at each weekly time. The area of hyper-fluorescence due to stroke damage is apparent, with its surface area decreasing over time. (E) Neuron Ca2+ traces (ΔF) for 100 neurons from one example animal pre-stroke (left panel) and 1 week post-stroke (right panel) in the peri-infarct somatosensory imaging region. Lower line graphs demonstrate the tracked movement (mm/s) within the floating homecage corresponding to the Ca2+ traces above.

Photothrombotic stroke results in forelimb representation remapping onto adjacent areas of cortex and altered sensory-evoked widefield Ca2+ response properties at 8 weeks post-stroke.

(A) Pseudocolored (%DeltaFoF) montages of representative cFL and cHL responses in S1 of an anesthetized animal pre-stroke and 8 week post-stroke resulting from oscillatory stimulation (1s, 100 Hz) of the cFL and cHL, respectively. At the 8 week post-stroke time, the cFL map has remapped posterior to its pre-stroke location into the area lateral to the cHL. (B) Mean peak forelimb Ca2+ transient response intensity (%DeltaFoF), measured from the thresholded cortical map area of the cFL. A significant interaction between time and stroke group was found, with post-hoc tests showing a significant decrease in the peak cFL response at 8-week post-stroke compared to pre-stroke (P = 0.0143). Mixed Effects Model, Time F(1, 12) = 0.8570, P = 0.3728; Stroke Group F(1, 12) = 0.2092, P = 0.6556. Interaction F (1, 12) = 8.632, P = 0.0124). (C) Thresholded cFL response area (mm2). A significant main effect of group and significant interaction was found, with the stroke group showing a significant enlargement in the cFL response area at 8 weeks post-stroke compared to pre-stroke (P = 0.0152), and a greater response area at 8 weeks in the stroke group compared to sham (P = 0.0036). Mixed Effects Model, Time F(1, 12) = 2.273, P = 0.1575. Group F(1, 12) = 7.328, P = 0.0191. Interaction F(1, 12) = 5.372, P = 0.0389). (D) The stroke group was found to have significantly greater total distance shifted for the forelimb map at the 8 week time compared to sham (P = 0.0002). Stroke N = 9, Sham N = 5. *p < 0.05; **p < 0.01; ***p < 0.001

Impaired performance on tapered beam task after stroke.

(A) Illustration of the tapered beam test with the three elements measured; left (contralesional) side slips, right side slips, and distance to first slip. (B) Mean number of left side slips. A significant interaction between time and stroke group was observed, with post-hoc tests showing a greater number of left side slips at 1 week in the stroke group relative to the sham group (P = 0.0435), and in the stroke group at 1 week relative to baseline (P = 0.0125). Mixed Effects Model, Time F (2.007, 38.13) = 2.178, P = 0.1270; Stroke Group F (1, 19) = 2.260; P = 0.1492. Interaction F(5, 95) = 2.914, P = 0.0172. (C) Mean number of right-side slips. No main effects or interactions were observed. (D) Normalized distance to first slip (as % of baseline). A trend towards a main effect of group was observed, with no trends found in post-hoc tests. Mixed Effects Model, Time F(3.135, 59.56) = 0.6447, P =0.5960.; Stroke Group F(1, 19) = 3.030, P = 0.0979; Interaction F(4, 76) = 1.242, P = 0.4596). Stroke N = 16, Sham N = 5. *p < 0.05; **p < 0.01; ***p < 0.001

Firing rate of neurons in the peri-infarct cortex correlates with performance on the tapered beam task.

(A) Example Ca2+ traces for 40 example neurons at each time point from the peri-infarct imaging region for one example animal in the stroke group. Significant Ca2+ transients are highlighted by red segments. A decrease in the number of Ca2+ transients is visually apparent at 1 week post-stroke. (B) In the peri-infarct imaging region, no main effects or interaction was seen in the normalized average firing rate during animal movement, however post-hoc analysis indicates a trend towards decreased normalized average firing rate at 1 week in the stroke group relative to sham (P = 0.0593). Mixed Effects Model, Time F (2.448, 29.99) = 0.9487, P = 0.4144; Stroke Group F (1, 13) = 0.04256, P = 0.8397; Interaction F (4, 49) = 1.852, P = 0.1339. A significant interaction was seen in the normalized average firing rate at rest (C), with post-hoc tests showing a trend towards a decrease in firing rate at rest between the stroke and sham groups at the 1 week post time (P = 0.0793, Mixed Effects Model, Time F (2.240, 27.44) = 1.048, P = 0.3712; Stroke Group F (1, 13) = 0.006936, P = 0.9349; Interaction F (4, 49) = 2.604, P = 0.0471). Normalized firing rate 1 week post-stroke was negatively correlated with the number of left slips in the tapered beam at the same 1 week time during both movement (R2 = 0.5326, P = 0.002) (D) and rest (R2 = 0.4346, P = 0.0075) (E). In the distal imaging region, no main effects or interaction were found in the normalized average firing rate for movement (F) or at rest (G). No correlation was observed between the normalized moving firing rate at 1 week in the distal region and the number of left slips on the tapered beam (R2 = 0.0637, P = 0.3286) (H). A statistical trend was observed towards a negative correlation between normalized firing rate at rest at 1 week with left slips on the tapered beam at the same 1 week time (R2 = 0.2320, P = 0.0503) (I). Peri-infarct region Stroke N = 11, Sham N = 4. Distal region Stroke N = 13, Sham N = 4. *p < 0.05; **p < 0.01; ***p < 0.001

Example cellular GCaMP6S Ca2+ fluorescence images and population functional connectivity plots.

(A-D) Ca2+ fluorescence images and functional connectivity plots from the peri-infarct imaging region and the distal imaging region for one example stroke group animal. Population functional connectivity plots demonstrate neurons as red dots, and the functional connection strength as lines with color and line width determined by the strength of the correlation between neurons. Notable cortical damage is visible in the upper portion of the 1 week post-stroke Ca2+ fluorescence image in the peri-infarct imaging region (A), which notably improves over the 2-4 week times. A notable loss of functional connections and loss in the number of connections with strong correlation (>0.3) is visible in the functional connectivity plots at 1 week post-stroke in the peri-infarct imaging region (B). In the distal imaging region post-stroke, less structural change is visible in the cellular Ca2+ fluorescence images (C), and less change to functional connectivity of the population over time (D). (E-H) Ca2+ fluorescence images and functional connectivity plots from the peri-infarct imaging region and the distal imaging region for one example sham group animal. Less structural changes within the cellular Ca2+ fluorescence images and in the functional connectivity plots are visually apparent relative to the example stroke animal.

Altered functional connectivity in the peri-infarct somatosensory cortex.

In the peri- infarct imaging region, a trend towards a main effect of group was found in the normalized average number of significant connections per neuron (A), with post-hoc tests showing a significant decrease at the 1 week post time in the stroke group relative to sham (P = 0.0440, Mixed Effects Model, Time F (4.132, 47.11) = 2.217, P = 0.1498; Stroke Group F (1, 13) = 4.505, P = 0.0551; Interaction F (4, 45) = 1.585, P = 0.1947). A significant main effect of group was found in the normalized number of significant connections per neurons with correlation greater than 0.3 (B), with post-hoc tests indicating a significant decrease at the 1 week post time in the stroke group relative to sham (P = 0.0363). (Time F (2.102, 22.60) = 0.3885, P = 0.6925. Group F (1, 13) = 6.423, P = 0.0262. Interaction F (4, 43) = 0.6512, P = 0.6292). (C) No main effects or interactions are seen in the connection density for the peri-infarct imaging region. (Time F (3.251, 40.31) = 0.3546, P = 0.8015. Group F (1, 13) = 1.257, P = 0.2824. Interaction F (5, 62) = 1.146, P = 0.3462). In the distal imaging region, no significant main effects or interactions are seen in the normalized number of significant connections per neuron (D) (Time F (3.173, 46.32) = 0.5674, P = 0.6487. Group F (1, 15) = 0.3172, P = 0.5816. Interaction F (5, 73) = 1.525, P = 0.1925), normalized amount of significant connections with correlation greater than 0.3 (E) (Time F (2.228, 32.86) = 2.196, P = 0.1224. Group F (1, 15) = 1.036, P = 0.3249. Interaction F (4, 59) = 0.7713, P = 0.5482), or in the connection density (F) (Time F (3.467, 52.01) = 1.299, P = 0.2841. Group F (1, 15) = 2.062, P = 0.1715. Interaction F (5, 75) = 1.251, P = 0.2942). Peri-infarct region Stroke N = 11, Sham N = 4. Distal region Stroke N = 13, Sham N = 4. *p < 0.05; **p < 0.01; ***p < 0.001

Color coded neural assembly plots depict altered neural assembly architecture after stroke in the peri-infarct region.

(A) Representative cellular Ca2+ fluorescence images with neural assemblies color coded and overlaid for each time depict a loss in the number of neural assemblies after stroke in the peri-infarct region, and a concurrent increase in the number of neurons for each remaining assembly. (B) Representative sham animal displays no visible change in the number of assemblies or number of neurons per assembly.

Neural assembly structure is disrupted in the peri-infarct region after stroke.

In the peri-infarct imaging region, a significant main effect of group was found for the number of assemblies (A) (Time F (2.883, 35.75) = 0.8231, P = 0.4857. Group F (1, 13) = 11.90, P = 0.0043. Interaction F (5, 62) = 1.496, P = 0.2040), with post-hoc tests showing a significant decrease in the number of assemblies at the 1 week and 3 week times for the stroke group compared to sham (P = 0.0114 & 0.0058, respectively). A statistical trend is also observed at the 2 week time for the stroke group compared to sham (P = 0.0589). A trend towards a main effect of group was found in the normalized number of neurons per assembly (B, Mixed Effects Model, Time F (2.361, 28.80) = 2.185, P = 0.1234; Stroke Group F (1, 13) = 3.694, P = 0.0768; Interaction F (5, 61) = 1.202, P = 0.3192), with post-hoc tests showing a significant increase in the normalized number of neurons per assembly at 1 week for the stroke group compared to sham (P = 0.0124). A trend towards a main effect of group was found in the average percent of the population per assembly (C, Time F (1.711, 21.22) = 1.738, P = 0.2023; Stroke Group F (1, 13) = 3.807, P = 0.0729. Interaction F (5, 62) = 1.765, P = 0.1333), with post-hoc tests showing a trend towards an increase in the stroke group relative to sham at the 1 week post time (P = 0.0863). No main effects or interaction was observed in the distal imaging region for number of assemblies (D), normalized number of neurons per assembly (E), or average percent of the population per assembly (F). Peri-infarct region Stroke N = 11, Sham N = 4. Distal region Stroke N = 13, Sham N = 4. *p < 0.05; **p < 0.01; ***p < 0.001

Correlation among neural assemblies is increased in the peri-infarct imaging region after stroke.

(A) Illustration of three example color coded neural assemblies from a neural population and their corresponding time series activations that were correlated. In the peri-infarct imaging region, a significant main effect of group was found in the correlation coefficient for significantly correlated assembly-assembly pairs (B, Mixed Effects Model, Time F (3.718, 42.39) = 2.014, P = 0.1141; Stroke Group F (1, 13) = 9.155, P = 0.0097. Interaction F (5, 57) = 2.348, P = 0.0524), with post-hoc tests indicating a significantly higher correlation at 2 week (P = 0.029) and 4 weeks (P =0.0253) in the stroke group relative to sham, and at 1 week (P = 0.0084) and 4 week (P = 0.0261) in the stroke group compared to pre-stroke. In the distal imaging region, no significant main effects or interaction was found for the correlation coefficient for significantly correlated assembly-assembly pairs (C). Peri-infarct region Stroke N = 11, Sham N = 4. Distal region Stroke N = 13, Sham N = 4. *p < 0.05; **p < 0.01; ***p < 0.001