Skip to main content
2024 BMES Annual Meeting Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Biomedical Imaging and Instrumentation

Biomedical Imaging and Instrumentation - Poster Session C

Poster F1 - Facilitating naturalistic mouse locomotive behavior during concurrent pupil tracking and wide-field optical imaging

Friday, October 25, 2024
10:00 AM - 11:00 AM EST
Location: Exhibit Hall E, F & G

Introduction:  Resting state functional connectivity (RSFC) analysis is a powerful methodology for assessing the brain's intrinsic functional organization, serving as a robust biomarker for various neurological diseases, such as Alzheimer’s and stroke. However, different brain states during task-free (i.e., resting-state) scanning can significantly influence RSFC estimates1,2. For example, RSFC estimates can vary greatly between across arousal states, such as wakefulness and sleep 2,3. Furthermore, emerging evidence suggests the presence of even more granular and continuously changing states, contributing to the complexity of interpreting RSFC2-4. These changes in arousal states may partly underlie/modulated large-scale, infraslow neuronal activity associated with RSFC measures3,4,5. The recognition of these fast state fluctuations has been linked to changes in pupil size5 and spontaneous behaviors (e.g., locomotion)6, making pupil and behavior monitoring critical for the accurate interpretation of RSFC5.

Wide-field optical imaging (WFOI) of cortical neuronal and hemodynamic activity facilitates RSFC analysis in awake, head-fixed mice. Although different groups1,8,9 have attempted to integrate pupil and behavior monitoring while neuroimaging, the experimental set-up often impedes naturalistic movements (e.g., squashed and unable to run, paw slips from poor grip, etc.), leading to potential stress and agitation. Here, we developed a system for concurrent recordings of WFOI, locomotion, and pupil tracking; a camera monitors pupil dynamics while an in-house treadmill consisting of a cylindrical wheel and a rotary encoder is used to assess spontaneous locomotion. Together with simple post-processing and deep-learning techniques, this system concurrently records pupil dynamics and locomotion in a manner more consistent with naturalistic locomotion and spontaneous behavior.

Materials and

Methods:  Two heterozygous Thy1-jRGECO1a mice underwent cranial windowing surgery for longitudinal WFOI (as described previously7). After acclimation, twelve 10-minute videos of task-free pupil dynamics were recorded under varying camera positions and lighting to ensure the deep-learning algorithm's generalizability. DeepLabCut (DLC) pose estimation was trained to track eight points around the pupil and one point in the center. Four randomly selected videos were K-meaned and the centroids of the ten largest clusters were manually labeled. After training, the algorithm's accuracy was evaluated, with outlier points manually relabeled, and the training dataset was appended, and the model was retrained. This process was repeated three times to achieve generalizability. Labeled points within each frame were then fitted to an ellipse using least squares approximation.

An in-house treadmill was designed to measure locomotive speed with an 8-by-4-inch cylindrical Styrofoam wheel mounted to a Taiss Rotary Encoder (600 pulses/revolution). A custom-built horizontal axle and encoder mount were created using Autodesk Fusion 360 (Figure 4). The rotary encoder's raw phase quadrature output was converted to speed using an Arduino Uno and sent to the same WFOI system data acquisition card. Speed and direction were encoded as pulse-width modulation through Arduino Uno digital output pins (PWM; Arduino, analogWrite, 490 Hz). An RC filter (fc = 7.985 Hz) converted PWM into analog signals. Mice were positioned on the wheel with a 30-degree downward head tilt at a height of 1.5’’. Mice were acclimated for 30 minutes over three days, and three 1-minute recordings extracted under task-free conditions.

Results, Conclusions, and Discussions: The pupil tracking algorithm yielded effective tracking, with labeled points regularly bordering the pupil edges. The first generation yielded inaccurate tracking (Figure 1A,B), but upon re-training, generalizability was achieved (Figure 1 C, D). Post-processing of pupil tracking displayed pupil area fluctuations (Figure 2). Our treadmill successfully facilitated naturalistic locomotion during head-fixation. Our velocity encoding algorithm demonstrates accurate quantification of locomotion speed and direction (Figure 3).

Design improvement is still needed for the current system. Although locomotion was achieved, it was consistently a slow walking behavior, which deviated from observations during unrestrained locomotion. Therefore, treadmill wheel size and head-fixation location (e.g., height) require systematic adjustments and assessments. In addition, other behavioral monitoring could be incorporated, for example, a camera facing the front of the mouse that assesses whisking. With the proper ensemble of hardware and software to track pupil and behavior, the interpretive power of RSFC as assessed by WFOI will be greatly improved.

Acknowledgements (Optional):  We would like to acknowledge the support of the American Heart Association (AHA Award #906519) and the WashU Cardiovascular Summer Research (CardS) program. I would like to thank Dr. Adam Bauer, Evan Morris, Jonah Padawer-Curry, and Nischal Khanal for their mentorship and guidance during this project.

References
[1] Barkus, C., et al. Journal of Neuroscience Methods 2022, 381, 109705
[2] Shahsavarani, S., et al. Neuron 2023,42,112527
[3] Raut, R., et al. Science Advances 2021, 7,30
[4] Gu, Y., et al. Frontiers in Neuroscience 2019, 13, 1662-4530
[5] Reimer, J., et al. Neuron 2014, 84, 355-362
[6] Syeda, A., et al. Nature Neuroscience 2024, 27, 187-195
[7] White, B., Bauer, A., et al. Plos One, 2011, 6, e16322
[8] West, S , A., et al. Cerebral Cortex 2022, 32, 2668–2687
[9] Dombeck, D. A., et al. Neuron 2007, 56, 43-57