Principal component analysis suggests multiple dimensions of memory inhibition that are differentially affected by age

BACKGROUND: Cognitive inhibition is among the executive functions that decline early in the course of normal aging. Failures to be able to inhibit irrelevant information from memory may represent an essential factor of age-associated memory impairment. While a variety of elaborate behavioral tasks have been developed that presumably all index memory inhibition, the extent to which these different tasks measure the same underlying cognitive construct that declines with age has not been well explored. METHODS: In the current study, 100 and 75 cognitively healthy younger (n = 71; age = 30.7 ± 5.4 years, 56.7% female) and older (n = 104, age = 69.3 ± 5.9 years, 66.2% female) adults with equivalent educational attainment performed three computer-based memory inhibition tasks: the Retrieval Induced Forgetting task, the Suppress task, and the Directed Forgetting task. We conducted a principal component analysis using scores derived from different components of these tasks to explore whether and how the tasks relate to one another. We further investigated how age, sex and education, along with, in a subsample of the participants, a neuropsychological measure of episodic memory, impacted both the task scores individually, and the principal components derived from the exploratory analysis. RESULTS: We identified 3 distinct sources of variability which represent potentially independent cognitive processes: memory retrieval facilitation, and two memory inhibition processes that distinguished themselves by the degree of volitional initiation of memory suppression. Only the memory retrieval component correlated with a neuropsychologically-derived episodic memory score, and both memory inhibition principal components were age dependent. CONCLUSION: Our findings provide support for a distinction in memory suppression processes between those ‘instructed’ to be performed and those which happen without explicit instruction. This distinction adds nuance to the dichotomous classification of controlled vs. automatic inhibitory mechanisms, which have been shown in previous work to vary as a function of the degree of frontal involvement. Our findings further demonstrate that while both of these measures of inhibition were affected by age, the episodic memory component was not, suggesting that inhibitory impairments may precede memory deficits in healthy aging.