"Automated Techniques to Model Brain and Behavior Connections in the Mouse Cortex: a Focus on Stroke"

Part of the Progress in Neuroscience Seminar (PINS) series Timothy H. Murphy Ph.D. Professor University of British Columbia Abstract The central focus of my lab is in understanding how cortical activity flow impacts normal brain function and diseases of the nervous system such as stroke. My laboratory contributes to understanding how a mouse cortex adapts after stroke, resulting in remapping of brain function from damaged to surviving areas using mouse models. The lab has developed new imaging and optogenetic tools that have parallels to human brain imaging and stimulation tools. We demonstrate that recovery after stroke employs circuits that are functionally related and structurally linked to those lost. Surprisingly, recovery not only involves functionally related circuits, but also network-wide changes in connection weights. To facilitate the study of stroke recovery we develop homecage brain-activity assessment tools for mice that are automatically followed in their own cages during short periods of restraint. Mouse head-fixed behavior coupled with functional imaging has become a powerful technique in rodent systems neuroscience. However, training mice can be time-consuming and is potentially stressful for animals. Here we report a fully automated, open-source, self-initiated head-fixation system for mesoscopic functional imaging in mice (Murphy et al. 2016 Nature Commun. and Elife 2020). The system supports 8 mice at a time and requires minimal investigator intervention. Using genetically encoded calcium indicator transgenic mice, we longitudinally monitor cortical functional connectivity up to 24 h per day. To accompany high-throughput assessment of brain activity, we develop a 3D synthetic animated mouse using animation and semi-random, joint-constrained movements to generate synthetic behavioral data for advanced model training. Image-domain translation produced realistic synthetic videos used to train 2D and 3D pose estimation that may facilitate automated etiological classification. We also integrate real-time behavioral assessment using markerless tracking of mouse body parts. The lab is currently integrating these tools to provide automated assessment of mouse stroke recovery within homecages. Publications Balbi M., Xiao D, Vega MJ, Hu H, Vanni M, Bernier L-P, LeDue J, MacVicar BA, Murphy TH (2021) Gamma frequency activation of inhibitory neurons in the acute phase after stroke attenuates vascular and behavioral dysfunction. Cell Reports 34, 108696Bolaños LA, Xiao D, Ford NL, LeDue JM, Gupta PK, Doebeli C, Hu H, Rhodin H, and Murphy TH (2021) A 3D virtual mouse generates synthetic training data for behavioral analysis. Nature Methods 2021 Apr;18(4):378-381. doi: 10.1038/s41592-021-01103-9.Murphy TH, Boyd JD, Michelson N, Fong T, Bolaños LA, Bierbrauer D, Siu T, Gupta PK, Balbi M, Bolaños F, Vanni M, and LeDue JM (2020) Automated task training and longitudinal monitoring of mouse mesoscale cortical circuits using homecages. eLife 2020 May 15;9:e55964. doi: 10.7554/eLife.55964.