Statistical learning is how humans extract regularities from the environment that are not explicitly taught. In other words, it is the explanation for how humans take meaningful information from the world without ever learning how. Rather than being taught, these regularities are often picked up over-time without a learner’s effort or awareness. Statistical learning is fundamental to many important regularities in our environment such as visual perception and language structure. So, how does this complicated learning work in children’s brains that are still developing?
Research on statistical learning in adults has shown that the left inferior frontal and superior temporal cortical regions have been implicated along with the hippocampus and the basal ganglia. Statistical learning showed a consistent lateralization of these effects. The prefrontal and temporal regions are most consistently implicated in the left hemisphere, especially for auditory paradigms. The caudate and hippocampus are most consistently observed on the right, but these subcortical findings tended to be more bilateral.
Previous studies on learning in young children showed that very young infants could learn more than we imagined them to. Infants treated strings of syllables that regularly occurred in sequence differently from strings that did not, after just 2 minutes of being exposed to an artificial language. This discovery concluded that infants are able to learn from passive exposure.
Prior to the present study, only two studies had even explored the abilities of children to statistically learn. One studied the functional ability for children to statistically learn and found increased recruitment of temporal cortices and greater activity in the left inferior frontal cortex for visually presented statistically regular relative to random items. The other study focused on the structural ability for children to statistically learn and found that the volume of the hippocampus predicts statistical learning performance across both children and adults. This concluded that previous studies have shown no difference in statistical learning outcomes between children and adults.
The present study by Amy S. Finn, Maria Kharitonova, Natalie Holtby, and Margaret Sheridan (2019) explored the relationship between statistical learning and the thickness and volume of structures across the brain that have been implicated in statistical learning and memory, focusing on the hippocampus and the caudate. The study focused on children aged 5-8.5 years old. This age period is when the primary neural substrates of interest are going through substantial functional and structural change.
This study discovered that the thickness of the left inferior frontal cortex and volume of the right hippocampus predict statistical learning ability in young children. This suggests that the frontal cortex is involved in statistical learning during childhood. The thickness or volume of these regions did not change over time with age. This is likely due to the tight age range (5-8 and a half years) measured in this study. On the other hand, the relationship between learning and the right hippocampus did change with age. Older children with smaller hippocampal performed better compared to the other children. This difference suggests that individual differences in hippocampal volume become more significant with age.
The findings of Amy S. Finn et al. (2019) concluded that because of hippocampal and left inferior prefrontal structure during childhood, children are capable of statistically learning. This is possible due to multiple neural structures that are more broadly implicated in learning and memory systems, especially in the hippocampus.
Finn, A. S., Kharitonova, M., Holtby, N., & Sheridan, M. A. (2019). Prefrontal and hippocampal structure predict statistical learning ability in early childhood. Journal of Cognitive Neuroscience, 31(1), 126–137.
