Colette Gazonas-Week 5-Shadlen Lab, Columbia University

During my fifth week in the lab I finished piloting the new mask paradigm we have been working on. I reached 40 percent as my minimum  success rate for each window, but the data was not as promising as anticipated so I was asked to increase this success rate to 80 percent. The grad student I work with analyzed this data and determined that there was a lack of symmetry within the graphs of psychometric curves folded one on top of each other. Therefore, he decided that we would need to define when a stimulus is difficult and when it is not difficult for each of the unfolded conditions which would require more trials.

I ran this paradigm about 50 more times this week which provided Yul with what he considered promising evidence in support of our hypothesis. Before this paradigm can be tested on naive subjects Yul said he would need to perform a little more data analysis but that he is optimistic about the results. By our next meeting he will have completed the analysis and will share with You-Nah and me what he has discovered by further investigating  some of his findings. We hope that this paradigm yields more reliable results than the less complex ones we previously ran in the lab and that we will be able to begin running it on naive subjects soon.

This week I also worked on completing a rough draft of my poster. Yul assigned me a different section to complete each night and by the end of the week I sent him a rough draft of the poster with everything organized on the template.

Over the weekend he got back to me with revisions. He suggested that in order to make my poster more approachable, I should save the words, and try to add a minimum of two more figures: one that illustrates the experimental procedure and another that illustrates the diffusion-to-bound model. Also, he noted that in one of the graphs I created in the the results section, the fit didn't seem right. His argument was that the curve doesn't overlap the data points from the actual data and suggested I may may need to check it.

By our next meeting on Tuesday I will have sent him a revised copy of my poster with less writing and more visuals. I will also have edited my matlab code for the graph of the predicted data so that the curve overlaps the observed data. I will do this three times, one for each of the three windows of stimulus duration and will plot all three curves on the same graph so they can easily be compared. Yul and I hope that my revised poster will serve as my final copy and that this will be done and approved by Tuesday night.