you are great

Hey there Mohamed -- unfortunately the audio issue stemmed from an audio channel not being recorded to begin with, so I have no way of restoring it. If there's a specific part of the video you'd like an explanation or any guidance for I can try explaining though.

@@KevinRblueneuronnet Thanks a lot, Kevin! I struggle a little in figure 4 in the paper, I didn't grasp the meaning. I haven't gone to the methods section yet, but when I do are you okay to help me counter something there? I was thrilled when I found the video to cover that part

@@MohamedAbdo-yp7pf -- Figure 4 is just comparing PSID to two other methods, NDM and RM. Specifically... 4a shows the number of state dimension each method needed to achieve the best decoding possible for each recording channel (the bar and number are the median, the box edges shows the 25th and 75th percentiles, the whiskers represent the minimum and maximum values that aren't considered outliers, and the asterisks indicate statistical significance). As you can see, PSID does much better. 4b compares the cross-validated Pearson Correlation Coefficients (used as a measure of decoding ability) between the predicted behavior and the true behavior in different regions (specifically the prefrontal cortex, the dorsal premotor cortex, the ventral premotor cortex, and the primary motor cortex). The square colors indicate the CC (i.e. yellow regions were decoded better, while blue regions were decoded worse), and the different colored asterisks tell us whether PSID outperformed NDM (if red) or RM (if black). 4c and 4d are the same things, only for a different monkey with bilateral coverage. Sure, feel free to reply with any other questions you have. Though you might want to check Omid Sani's talk about this paper at NMC first: kzbin.info/www/bejne/j6KUhJuOoNCKr7s

@@KevinRblueneuronnet Thank you so much Kevin! You have been a great help indeed