In the pilot study we will do, the corner test will be used as the primary outcome. The concern I have is that the corner test AI results presented in our weekly meeting did not show the difference between the early time point and the late time point. Functional deficits were supposed to improve at 28 days. However, the corner test did not catch this change, which makes me worry about whether this test will not catch the treatment efficacy if the drug effect is not strong.
The principle of the corner test is to evaluate the response of animals to the vibrissae stimulation. It is the same as the vibrissae touch in the 48-point score. The right MCAO results in a response loss on the left vibrissae. So, they turned to the left on the original paper. This test is good to be used in the distal MCAO. In the filament MCAO, the striatum area is also damaged, which results in a turning behavior to the ipsilateral side. Such circling behavior interrupts the corner test performance.
Some groups used the corner test to evaluate the circling behavior in the stroke experiment, which is not what was tested in the originally designed corner test, for example, the stroke mouse did not respond to the touch of the left vibrissae and however, it turned to the right. The circling behavior can be observed in the open field and also included in the 48-point score.
I strongly suggest to analyze some SPAN1 data, including the corner test scores, the cortical lesion sizes and striatum lesion size at 48 hours of MRI. We need to determie if the scores match the lesion size (total lesion, cortical lesion only, and striatum lesion only) before the pilot study begins.
I agree the corner test is…
I agree the corner test is problematic for various reasons. We need more than just 1 test to hang our SPAN hats on, and not put all our eggs in one basket. Given that the trial statistics need a single primary endpoint, a composite score of sorts must be developed. We tried the 48-point test, but it is really lengthy in our hands, and some aspects were difficult to score accurately.
There is no reason not to combine and modify existing behavior tests, but it requires validation. Since SC meets once a week, and sometimes not even that, I am not sure whether this task can be accomplished before the pilot starts. I have suggested a Behavioral Assessment Committee actively focus on this and spend much more time researching and debating. They can then present to the SC, educate us, and help make an informed decision.
Cenk
I worry that the corner test…
I worry that the corner test will not detect the efficacy of new compounds on cortical lesions.
Once we agree on the…
Once we agree on the behavior tests we want to do, it is trivial for stats to create a composite score to feed into the MAMS procedure.
I thought the neurological…
I thought the neurological examination would be acceptable if it is similar to what is used in humans. When video recording is required, it becomes challenging and complicated. Could we skin the video recording and let each site score the stroke animals by two people and use the average score? The 48-point scoring system includes a complete set of motor and sensory deficits; for example, front limb symmetry is similar to the cylinder test, and vibrissae touch is identical to the corner test.
If this scoring system is not selected for the pilot study, we should include another test to detect post-stroke sensory deficits.
We also think the core area will be less affected by drug treatment. That is why we like to examine other body parts, not just the front limb.
rearing/standing before turn
The Corner Test SOP states that “the mouse may rear forward and upward, stand up, and/or then turn back to face the open end”. Indeed, in many papers, only turns with rearing/standing are counted valid in this test. I guess that this kind of behavior chain may indicate that the turn is triggered by vibrissal touch. Not sure whether the AI software also takes this into account.