Extended Data Fig. 7. Model free analysis confirms the minimal modulation by movements.
Trials were divided into two groups (‘with movement’ versus ‘without movement’ trials) depending on whether the animals moved on these trials. The movement was quantified as the average |pixel-difference| across consecutive video frames, and we used this metric to classify trials (see Methods). The average stimulus-driven responses across all units for with and without movement trials are nearly identical in both monkeys (a, b, left). When the average responses are aligned on the onset of the movement (a, b right) if movements were present (blue), or for matched stimulus conditions without movement (orange), the responses on epochs with movement are very similar to corresponding epochs those without movement. It indicates that there was no systematic modulatory effect relative to the movement onset across the population of units. (c) We also computed a movement modulation index (MI, see Methods) for each unit. On average, the MI in each area was close to 0 (mean ± standard deviation; MI = -0.009 ± 0.029 p = 10−5; −0.011 ± 0.036, p = 10−5, −0.008 ± 0.032, p = 10−4; for V1, V2, V3/V3A, respectively; two-sided signrank test for deviations from 0, corrected for multiple comparisons); identical scale of the x-axis as in Fig. 3 to facilitate the comparison of the effect sizes. This model-free analysis therefore confirms our model-based analysis and shows that any modulation by the animals’ own movements in these visual areas is very small.