MT-1 repeated stimulus

Hi, I and a partner are using the MT-1 dataset for a class project and we are unsure if any of the provided cells had a repeated stimulus. The paper says that for 20 of the recorded neurons, the stimulus was repeated every 5 seconds with the aperture centered on the cell's receptive field, and we assume one of those cells was used to make the raster in Figure 1B.

Thank you for your interest in the dataset! We are glad for this question, although I apologize that this aspect of the experiment data is not available. 

We had two practical reasons, as well as a conceptual reason for not including the repeated data -- if I am remembering correctly. Practically: (1) the repeated data was recorded for only a subset of neurons in the study; and (2) quite difficult to use given fixation breaks and other detailed structure issues in the dataset, particularly for trying to getting this dataset in shape for public posting. Conceptually, while *seeming* to offer a more interpretable measure of cross-validated performance, I would argue that the R^2 values one gets for repeats in visual neuroscience experiments will have deeper interpretability issues, as we explored in further papers from our lab: Cui et al, J Neuroscience 2016) and McFarland et al, J Neuroscience, 2016. [Briefly, this concerns both the amount of "unexplainable" variability in each neuron; and (2) what is explainable versus unexplainable in the context of other factors such as eye movements and aspects of uncontrolled brain state. As a result, we only used it to generate the example figure and provide a "calibration" the more sensitive/accurate (but less familiar) performance measures of cross-validated log-likelihood (LLx).