Multitasking Effects on Perception and Memory in Older Adults

Abstract

Performing multiple tasks in parallel is detrimental to performance, a phenomenon generically referred to as dual-task interference (DTi). Several variables can modulate DTi at the individual level, and increasing age is typically described as negatively affecting response costs. In this study, we investigated, in 252 healthy adults aged between 50 and 89 years, how age modulates the detrimental effect of DTi during the encoding of images. We combined a visual memory task and a sustained attention task (i.e., an auditory version of the continuous performance task, ACPT) in three separate blocks. In the first block, participants had to perform a four-alternative forced-choice recognition of previously memorized images without having attended to ACPT sounds during the encoding. In the second block, during memorization, participants had to press a response key when detecting the letter "X" within a stream of letters (Low Load). In the third block, they had to respond only when the letter "X" was preceded by the letter "A" (High Load). The results showed that overall performance linearly decreased with age. In contrast with our predictions, DTi was stable across different ages. Finally, using a cluster-based approach, we found that participants who paid the highest costs when dual-tasking also demonstrated, on a self-administered cognitive screening significantly lower scores than peers. These new types of tests, which ask for concurrent task performance, might become useful for detecting outlier performance that might anticipate or correlate with aging-related cognitive decline.

Keywords: aging; cognitive load; cost; dual-task; memory.

Figure A1

Memory reaction time as a function of Age and Cognitive load (No Load, LL, and HL). Bands represent 95% confidence intervals.

Figure A2

Years of schooling as a function of age.

Figure 1

Diagram of participants’ recruitment.

Figure 2

Frequency plot of the participants’ ages. Each bar represents the number of participants in each age group (5 years).

Figure 3

A representative sequence of the memory task. During the encoding phase (upper panel) 15 images were shown in sequence for 5 s each. Concurrently, a stream of 45 auditory stimuli (letters) was played with an onset asynchrony of 1666 ms. In the No-Load condition (NL), participants had to ignore the auditorily presented letters. In the Low-Load condition (LL), participants had to press a response key each time the letter “X” was presented. In the High-Load condition (HL), participants had to press a response key whenever the letter “X” was preceded by the letter “A”. During the testing phase (lower panel), a forced-choice image recognition was performed. One of the four images consisted of a previously presented (target) image (here image number 3, highlighted in green).

Figure 4

Performance in the memory task (memory score) as a function of Age (in years), reported separately for the three Cognitive Load conditions (No Load, LL, and HL). Bands represent 95% confidence intervals.

Figure 5

The ACPT score as a function of Age and Cognitive Load (ACPT-A: Low Load, ACPT-AX: High Load). Bands represent 95% confidence intervals.

Figure 6

Correlation between the mnestic dual-task interference DTi (No Load–HL) and ACPT DTi (LL–HL). The bands represent 95% confidence intervals. Positive values indicate that more errors were committed in the High Load condition than in the Low Load (ACPT) or No Load (memory task) conditions.

Figure 7

Graphical representation of participant dissimilarity calculated with a k-medoids clustering algorithm over three cost indices: the two cost indices from the memory test (No Load minus LL or HL for memory DTi) and the cost index from the ACPT (LL–HL for ACPT DTi). The cluster with low DTi includes 77 participants, the cluster with mid-DTi 88, and the cluster with high DTi 87.

Figure 8

Distribution of participants’ auto-GEMS scores in the three clusters: low-DTi, mid-DTi, and high-DTi.