Statistical Learning Mechanisms Revealed Through Older Adult Memory Performance

Statistical learning (SL) refers to behavioral benefits when identifying regularities in the environment. For example, in parsing language, ‘ing’ typically marks the end of a word, or in vision, the presentation of one image predicts the arrival of another. SL is considered implicit and nondeclarative because it does not rely on conscious, deliberate effort. The neural correlates of implicit learning are classically linked with dorsal striatal function and not with medial temporal lobe areas associated with episodic memory function. However, recent data from a patient with selective hippocampal damage found she showed no visual SL. A separate literature has examined whether SL is passive or whether attention is required. These observations are rekindling interest as to whether SL is a general mechanism for acquiring statistical regularities in the environment, as an implicit learning mechanism or whether it actually is an explicit memory function relying on the hippocampus. Existing data indicate that SL tasks are relatively preserved with age. This observation is consistent with the implicit view. In contrast, if SL is explicit, the prediction would be to observe age-related decline in SL performance, commensurate with age-related memory and attentional decline. Infant work has demonstrated that SL mechanisms are sensitive to co-occurring frequencies, and transitional probabilities between stimuli, such as when learning language. Transitional probability allows for segmentation of words from non-words in continuous speech streams, for example. Such sensitivities acquire structure from the environment. When transitional probability is held constant, structure can still be acquired, termed “community structure”. Currently, community structure is considered a higher-order form of learning, and is not synonymous with SL. Topology must be clarified to understand whether SL is an implicit or explicit mechanism (or both), and how other forms of learning such as community structure fit within this framework. Comparing older adult behavior on these paradigms along with their attentional ability will clarify whether these behaviors are relatively well maintained. The results showed that older adults continue to exhibit intact SL. The community structure task results hint towards retention in older adults. Demonstrating an intertwined relationship, sustained attention significantly correlated with SL behavior whereas divided attention did not. Collectively, investigating SL is meaningful in developing our understanding of learning processes throughout the lifetime. Routines and patterns play an important role in everyday life and provide important information about the environment