Statistical Learning Signals for Complex Visual Images in Macaque Early Visual Cortex

Abstract

Animals of several species, including primates, learn the statistical regularities of their environment. In particular, they learn the temporal regularities that occur in streams of visual images. Previous human neuroimaging studies reported discrepant effects of such statistical learning, ranging from stronger occipito-temporal activations for sequences in which image order was fixed, compared with sequences of randomly ordered images, to weaker activations for fixed-order sequences compared with sequences that violated the learned order. Several single-unit studies in macaque monkeys reported that after statistical learning of temporal regularities, inferior temporal (IT) neurons show reduced responses to learned fixed-order sequences of visual images compared with random or mispredicted sequences. However, it is unknown how other macaque brain areas respond to such temporal statistical regularities. To address this gap, we exposed rhesus monkeys (Macaca mulatta) to two types of sequences of complex images. The "regular" sequences consisted of a continuous stream of quartets, and within each quartet, the image order was fixed. The quartets themselves were displayed, uninterrupted, in a random order. The same monkeys were exposed to sequences of other images having a pseudorandomized order ("random" sequence). After exposure, both monkeys were scanned with functional MRI (fMRI) using a block design with three conditions: regular sequence, random sequence, and fixation-only blocks. A whole-brain analysis showed a reduced activation in mainly the occipito-temporal cortex for the regular compared to the random sequences. Marked response reductions for the regular sequence were observed in early extrastriate visual cortical areas, including area V2, despite the use of rather complex images of animals. These data suggest that statistical learning signals are already present in early visual areas of monkeys, even for complex visual images. These monkey fMRI data are in line with recent human fMRI studies that showed a reduced activation in early visual areas for predicted compared with mispredicted complex images.

Keywords: expectation; monkey fMRI; predictions; sequence learning; statistical learning; visual cortex.

FIGURE 1

Stimuli and functional MRI (fMRI) paradigm. (A) One group of images (e.g., images of the left panel) was employed as the regular sequence in one monkey and as the random sequence in the other monkey and vice versa. The images were presented in quartets (stippled green box in the left panel), resulting in five fixed quartets (rows) for the regular sequence. The random sequences were obtained by selecting for each presented quartet at random one of the five images at each position [column in (A), right panel] of the quartet. The dashed orange outlines indicate one such possible selection. In this example, the quartet would be “cow, eagle, dog, hen.” For more details, see “Materials and Methods.” (B) Mean of the images for each of the two groups. (C) Example of a run during scanning. Reg, regular sequence; Ran, random sequence; Fix, fixation only condition; Fix Start, fixation only period at the beginning of the run. The duration of each block is indicated. The order of the conditions was randomized across runs.

FIGURE 2

Stimulus activations: contrast random sequence – fixation only. Sagittal and coronal slices illustrating for each monkey and hemisphere the activations by the images compared to a condition in which only the fixation target was presented. The activations are presented on the MRI anatomy of each monkey. The vertical lines indicate where the slices were taken. The color scales indicate t values, thresholded at p = 0.05 family wise error rate. (A) M1, (B) M2.

FIGURE 3

Activations when contrasting the regular and the random sequences. Four representative slices are shown for each monkey using their own anatomy. The red activations correspond to the contrast random–regular sequence, while the blue ones correspond to the opposite contrast. (A) M1, (B) M2. The same conventions as in Figure 2.

FIGURE 4

Cortical activations shown on a probabilistic retinotopic map of the inflated F99 monkey brain. The thresholded retinotopic maps are indicated in light gray, using the definitions of retinotopic areas by Zhu and Vanduffel (2019). The thresholded cortical activations (p < 0.05 family wise error rate) corresponding to the contrast random–regular are shown in red–yellow; those of the opposite contrast in blue. (A) M1, (B) M2.

FIGURE 5

Eye movement metrics of the pre-scanning exposure phase (last eight exposure sessions). The means and 95% confidence intervals for the regular (Reg), random (Ran), and deviant (Dev) quartets are plotted for monkey M1 and M2. Std dev, standard deviation.

FIGURE 6

Eye movement metrics during functional MRI (fMRI) for the same runs that were employed for the fMRI analysis. The standard deviations of the eye position outside the blink epochs were computed per block for the horizontal (top) and the vertical dimensions separately. The means and 95% confidence intervals for the regular (Reg) and random (Ran) sequences are plotted for monkey M1 and M2.