For a quick overview for those just joining, I am examining the results of an experiment on latent inhibition using a video-game method by developing a tool to rapidly view and group the eye-tracking data. Some of the data I’ve already described.
After a quick fix of the program I can now collect and export the “hit” data, which are simply counts of the number of gazes recorded in the area of interest.
I do not bother calculating “fixations” which are defined by gaze points being X distance apart within a Y period of time. It seems to me that X and Y could be somewhat arbitrary. Also, the data were recorded at 60hz, so most of what is collected represents data from “fixations” anyway as saccades are generally fast enough that 60hz will miss most.
Interestingly, the participant is able to see the entire screen, regardless of where they fixate. Some subjects were observed charging the weapon when the CS came on, sometimes seconds before they ever looked directly at the CS. Thus, even a saccade through the area would be more than sufficient to trigger a response.
A final reason that I favor the raw data is that, at 60hz, each sample represents a snapshot in time. Thus, time is represented in each data point. If, for example 30 data points were recorded then 30/60 = .5 seconds of time spent looking in that area.
The data above represent the number of gaze hits within the area of the sensor CS during each second of conditioning. These data are noisy: Nothing is statistically significant, though there are several ps < 0.08. Tighter procedures and more subjects will clean that problem up nicely. So, during conditioning it will be clear that the latent inhibition group, which was just pre-exposed to the sensor, is not looking at it very much. In contrast, the Control group begins looking at it, and attention decreases over trials (as Hall & Pearce would predict).
The data below are from the five trials of extinction, second by second during the entire 20-s of the CS.
There are significant effects of Group, trials, seconds and Trials X Group in these data.
There are many interesting things. Notice the control group, especially trial 1, that their vision data mirror their behavioral data. As they begin to respond more, they look at the CS less. After about 5-6 seconds, when the US should have arrived, they begin to respond less, but look at the CS more. After this first extinction trial attention to the CS goes up as shown on the next trial, and then declines once more. These data are again consistent with Hall & Pearce.
In neither the conditioning, nor the extinction data was attention to the CS very evident in the latent-inhibition group.
Next: I shall analyze the pupil data I presented earlier.