How a caregiver perceives and responds to infant cues will determine whether the infant receives the care it needs to survive. In mice, pup vocalizations evoke complex behavioral responses, and thus likely evoke coordinated activity across a network of brain regions. We want to identify this network in mice of different maternal experience backgrounds, and awake mouse FMRI is a great tool for that.
As awake mouse FMRI becomes more widely used, exciting research is emerging that investigates the neural correlates of goal directed behaviors. These behaviors are typically assessed using tools like beam breakers to the detect leaks and pressure sensors to detect movements. Our protocol allows us to assess spontaneous behavioral responses, even in the technically challenging MRI environment.
Our goal is to eventually use measures of spontaneous behavior to better understand the brain-wide activity we acquired with FMRI. Our prior work revealed that the auditory cortices of mom and virgin animals respond differently to pup calls. In moms, their neural responses are timelocked to the cry of a pup.
In virgins, they're not. However, as a virgin learns to express maternal behavior, the auditory centers change to reflect this new skill. Our prior studies focused on auditory cortex in anesthetized mice.
This study pushes beyond the boundaries of the auditory cortex, examining how the entire brain and behavior changes in awake behaving mice as an animal learns to become a mom. To begin, download MATLAB from the MathWorks website. To add the image processing toolbox to MATLAB add-ons, click on manage add-ons, get add-ons, and then search for and add them.
Similarly, add the Computer Vision Toolbox to the MATLAB add-ons. After opening script one of the analysis pipeline, edit sections one, two, three, four, five, and six of the script to fit data structures where indicated. Run section one.
Run section two of the script to pull up the first frame of each video. Use the mouse to click once on the first two points, then double click on the third point. After selecting the last points of the video, press enter.
Run section three and section four. Set the flag aligned to zero and run section five. Then set the flag aligned to one and run section five again.
Now run section six and compare the results side by side. Next, open script two and edit section one of the script to fit data structures. Then run section one.
Edit section two of the script to fit data structures and run section two. Use the resulting image to see where optical flow fluctuates most dramatically in the analyzed videos. Edit section three of the script to fit data structures and analysis needs.
To analyze the full video FOV, set the flag select ROI and provide ROI to zero. Then run section three. To analyze a new ROI, set the flag select ROI to one and provide ROI to zero.
Then edit newCoordsName and run section three. Select a new ROI in the displayed example frame by clicking and dragging. For analyzing previously drawn ROI, set the flag select ROI to zero and provide ROI to one.
Then edit inputCoords to provide a predetermined set of coordinates and run section three. Open script three and edit sections one to three of the script to fit data structures and analysis needs. In section one, set the analysis options flags ZSC, BLRM, BLZSC, demeanPerTrial, and applyLPfilter to zero or one.
Set LPfilter to a number representing the desired lowpass filter in hertz. Then run sections one to three. Open script four and edit section one of the script to fit data structures and analysis needs.
Run section one. Open script five and edit section one of the script to fit data structures and analysis. Then run section one and section two.
Change the group, myTitle, and threshT variables in section three inputs to reflect the group analysis desired. Run section three for each group of interest. Then adjust the range of the C axis and the color bar limits as desired for visualization.
Change the mag, myTitle, and threshT variables in section four inputs to reflect the group comparison desired. Run section four for each group comparison of interest, then adjust the range of Z axis and the color bar limits as desired for visualization. Optical flow vectors indicated regions of animal movement, especially around the snout and paws.
The standard deviation of optical flow magnitude identified areas with high and low movement fluctuations. Mothers showed higher movement responses to pup calls compared to virgins, but not to pure tones. Principle component analysis revealed that mothers had more movement in the nose area compared to virgins during both stimuli presentations.
Optical flow quantification of a validation dataset revealed significant differences in behavioral responses to high versus low water rewards. These results further suggest that the presented video analysis pipeline can be used to capture meaningful differences in animal behavior across conditions.
