In a recent study published in the prestigious journal Current Biology, researchers from the Universities of Pisa, Sydney, Florence and Salento have shown that statistical learning – that is, where we acquire information in a completely automatic and unconscious way – can even be traced back to one of our simplest and most unconscious reactions: the constriction or dilation of the pupil of our eyes, evoked by the sight of an image.
“This study shows that our visual system is sensitive to the statistical regularities of our environment even when we are not consciously aware of them,” says Paola Binda, professor at the University of Pisa and lead author of the study. “Pupil diameter remains a valuable source of information on how our sensory and cognitive systems work: a real window into the mind and its learning abilities.”
The study starts from the premise that a large amount of the information on which our behaviour is based is learned spontaneously and unconsciously. Just think of language acquisition: we are able to distinguish words from the sound produced by the speaker, even though it is continuous and lacks clear pauses to mark the end of one word and the beginning of the next: “To learn, we don’t need instructions or directions,” continues Paola Binda, “we can do it from the first weeks of life, simply by listening to the sounds of our language. Perhaps, this form of ‘statistical learning’ is important for extracting meaning from all sensory signals, not only auditory but also visual, tactile, etc.
For their research, the scientists showed the patients images of sets of seemingly random bars (link to video). Their temporal succession was very fast and regulated by a simple statistical structure: each image containing 24 bars was followed by one with 6 bars, 2 bars were followed by 12 bars, and so on, creating fixed numerosity pairs. Due to the rapid succession of images and the variable arrangement of elements, this temporal structure was imperceptible. Nevertheless, the pupil diameter oscillated systematically in response to the repetition of pairs (whereas no oscillation was observed in a control experiment in which the same images were presented in random order).
“Thanks to this innovative methodology, it is possible to monitor the evolution of complex brain processes in an indirect and non-invasive way”, concludes Binda. “In the long term, this type of research could provide us with new tools to characterise inter-individual differences and dysfunctions in learning.”