Colette Gazonas-Week 2-Shadlen Lab, Columbia University

The work I did in the lab this week was very similar to what I did last week. I spent the majority of my hours in the lab running a series of paradigms which implemented a mask. The purpose of this new paradigm is to test whether or not we can retain information from a very short stimulus even after the stimulus ends and is replaced with a mask. In this case, the small dots of varying direction and color are replaced with a constant array of larger dots which are evenly distributed between blue and yellow and which do not follow a certain direction of motion.

This test consists of three different windows. The order in which each window of x amount of trials appears  is random and therefore varies from trial to trail. In the first window the stimulus is present for 80 milliseconds, the second window shows a stimulus for 160 milliseconds, and the third and last window shows the stimulus for 320 milliseconds. In order for a test to be considered a success, and statistically significant, the success rate for each of the three windows much exceed 90%. Note that for all of these paradigms the success rate is measured by timing accuracy, not by whether or not you choose the correct target (right or left, blue or yellow). This means that the subject must initiate eye movement directly after the mask disappears and they must arrive at a target within a specified number of milliseconds. The reason success is measured in this way is because it ensures that the subject understands the paradigm and what they are being asked to do. By making their decision within the timeframe it indicates that they understand the task at hand and therefore verifies that the decisions they are making are in response to the stimulus, not random eye movements.




The machine I use to run the paradigms which focus on the detection of eye movement

I ran this test in two variations, one asking me to dictate which color was dominant and the other asking me to indicate which direction the dots were moving. By the end of the first half of my third week in the lab I hope to have enough successes in this paradigm to have sufficient data to analyze.

On Wednesday I attended my first journal club discussion. At the beginning of the week I was emailed the link to the article that would be discussed at the meeting. It was about the way we see and it provided evidence against the theory of remapping. During this lab meeting one of the lab members created a PowerPoint in which she summarized the big ideas of the article. The lab deciphered the graphs and diagrams provided in the article and shared their views and opinions about it. A lot of the terminology discussed during the meeting was very sophisticated and required that I have a solid background in neuroscience to understand the majority of it. This being said, I was unable to comprehend what was being discussed in total, but I was able to get the gist of it.

For my matlab assignment this week I was to generate a random set of data for motion coherence by using the formula logit(p)=B0 + (B1 - B2*Icor)*x. I did this by assigning random values to each beta and choosing either 0 or 1 as the Icor value. I plugged in [-1,-.5,-.25,0,.25,.5,1] for the x values in order to find logit(p) and from their solved directly for p. Then, I used glmfit to graph the data along a logistic curve.

Overall, my second week in the lab was a bit more challenging than my first because now that I was familiar with the software and paradigms the difficulty level on them was taken up a notch. Although I encountered some trouble figuring out matlab along the way, I was able to figure out a way to get the assignment finished which I was proud of. I hope by the third week in the lab I will be able to say I am very comfortable using matlab to generate and analyze data.