Figure 1. Visual stimuli and behavioral task.

(A) Schematic of the four kinds of texture discrimination tasks administered to the four groups of rats in our study. Each group had to discriminate unstructured binary textures containing white noise (example on the left) from structured binary textures containing specific types of local multipoint correlations among nearby pixels (i.e. 1-, 2-, 3-, or 4-point correlations; examples on the right). The textures were constructed to be as random as possible (maximum entropy), under the constraint that the strength of a given type of correlation matched a desired level. The strength of a correlation pattern was quantified by the value (intensity) of a corresponding statistic (see main text), which could range from 0 (white noise) to 1 (maximum possible amount of correlation). The examples shown here correspond to intensities of 0.85 (one- and two-point statistics) and 0.95 (three- and four-point statistics). (B) Schematic representation of a behavioral trial. Left and center: animals initiated the presentation of a stimulus by licking the central response port placed in front of them. This prompted the presentation of either a structured (top) or an unstructured (bottom) texture. Right: in order to receive the reward, animals had to lick either the left or right response port to report whether the stimulus contained the statistic (top) or the noise (bottom). Figure 1—figure supplement 1 shows the performances attained by four example rats (one per group) during the initial phase of the training (when the animals were required to discriminate the stimuli shown in A), as well as the progressively lower statistic intensity levels that these rats progressively learned to discriminate from white noise during the second phase of the experiment.

Figure 1—figure supplement 1. Learning curves of four example rats during the two initial training phases.

(A) Learning curves of four example animals belonging to the first batch of rats (see Materials and methods), each trained with one of the four texture statistics: one-point (red), two-point (blue), three-point (purple), and four-point (green). During this phase, animals had to discriminate textures containing a single high intensity level of the assigned statistic (see examples in Figure 1A, right) from white noise. Each curve shows the performance of a rat (i.e. the fraction of correct choices), as computed in consecutive blocks of 500 trials. The dashed line indicates the 65% criterion performance that rats of the first batch had to reach in order to be moved to the next experimental phase (see Materials and methods). As exemplified by these four animals, rats trained on different statistics required different lengths of time to learn the discrimination – animals trained on one- and two-point statistics reached and maintained a performance above criterion considerably earlier than those trained on three- and four-point. Specifically, considering only those rats that reached criterion performance, animals trained on one-point correlations required on average 3.5 ± 0.3 blocks of 500 trials (n = 9) to reach criterion; those trained on two-point correlations 5.3 ± 1.3 (n = 8); those trained on 3-point statistic 3.4 ± 1.6 (n = 5); and those trained on four-point correlations 11.8 ± 1.6 (n = 10). (B) Minimal intensity levels of the statistics that the same animals shown in A were able to discriminate from white noise across consecutive sessions (colored dots) of the second phase of the experiment, when the rats were exposed to progressively lower intensities through a staircase procedure (see Materials and methods). Sessions where the number of trials performed by the animals was below 50 are not shown in these graphs.