Tactile memory impairments in younger and older adults

Abstract

Human tactile memory allows us to remember and retrieve the multitude of somatosensory experiences we undergo in everyday life. An unsolved question is how tactile memory mechanisms change with increasing age. We here use the ability to remember fine-grained tactile patterns passively presented to the fingertip to investigate age-related changes in tactile memory performance. In experiment 1, we varied the degree of similarity between one learned and several new tactile patterns to test on age-related changes in the "uniqueness" of a stored tactile memory trace. In experiment 2, we varied the degree of stimulus completeness of both known and new tactile patterns to test on age-related changes in the weighting between known and novel tactile information. Results reveal that older adults show only weak impairments in both precision and bias of tactile memories, however, they show specific deficits in reaching peak performance > 85% in both experiments. In addition, both younger and older adults show a pattern completion bias for touch, indicating a higher weighting of known compared to new information. These results allow us to develop new models on how younger and older adults store and recall tactile experiences of the past, and how this influences their everyday behavior.

Keywords: Consolidation; Episodic memory; Haptic; Hippocampus; Long-term memory.

Figure 1

Stimuli and Experimental Setup. (A) Overview of experiment 2: The three ‘learned’ tactile patterns were learned during learning phases 1 and 2. The three ‘new’ tactile patterns were presented in the experimental phase and had to be dissociated from the three learned patterns. Black dots indicate the elevated and vibrating pins of each tactile pattern, white dots indicate the pins that were not elevated and did not vibrate. Both ‘learned’ and ‘new’ tactile patterns were either presented in full (completeness level 100%) or at reduced completeness (completeness levels 87.5%, 75%, 62.5%, 50%). (B) Experimental trial of experiment 2: Participants were informed of the completeness level of the tactile pattern via a bar on screen. After tactile stimulation, they were offered a 4-alternative-forced choice response (‘pattern1’, ‘pattern2’, ‘pattern3’, ‘new’). A Piezo Stimulator with a 4 × 4 stimulation module was used to create the tactile patterns. (C) Overview experiment 1: The ‘reference’ tactile pattern was learned during learning phase 1. The seven ‘new’ tactile patterns differed in 1–7 pins (i.e., in 87.5–12.5%) from the ‘reference pattern’. Black dots indicate elevated and vibrating pins that remained the same as in the ‘reference pattern’, white dots indicate the pins that were not elevated and did not vibrate, gray dots indicate pins that were elevated and vibrated, but differed from the ‘reference pattern’. The same stimulator as in experiment 2 was used (shown in (B)).

Figure 2

Precision to differentiate one learned tactile pattern from several new tactile patterns with varying similarity in younger and older adults (experiment 1). (A) Mean accuracy to correctly recognize the ‘reference pattern’ and the ‘new patterns’ plotted for the different similarity levels for younger (blue) and older (red) adults (mean ± CI). 50% marks chance level. (B) Raincloud plots for data shown in (A), individual data shown as colored dots: younger adults = blue and older adults = red. Boxplots are drawn within the interquartile range (box), medians are shown as vertical lines, whiskers connect the minimum and the maximum with the lower and the upper quartiles. (C) Main effect of age on accuracy to correctly recognize the ‘reference pattern’ and the ‘new pattern’ (mean ± CI). (D) Effect sizes (Cohen’s d) for data shown in (B) plotted for the different similarity levels. Negative values indicate lower performance of older compared to younger adults, positive values indicate better performance of older compared to younger adults. (E, F) Accuracy to correctly recognize a ‘new pattern’ after goodness-of-fit estimates for each individual of younger and older adults (thin lines), and mean curve for younger and older adults (thick lines, 50% thresholds: younger = 3.15, older = 3.42, p = .467, Cohen’s d = 0.36, thin dotted lines indicate 50% correct performance).

Figure 3

Performance to discriminate learned from new tactile patterns at different completeness levels in younger and older adults (experiment 2). (A) Mean accuracy to correctly recognize the learned patterns’ and ‘new patterns’ and the ‘new patterns’ plotted for the different completeness levels for younger (blue) and older (red) adults (mean ± CI). Blue line and red line indicate mean values for younger and older adults across completeness levels. Chance level is 25% (black line). (B) Effect sizes (cohen’s d ± CI) of data shown in (A). (C) Boxplots of bias scores for younger and older adults plotted for the different completeness levels. Bias scores were calculated by subtracting the accuracy in the new conditions from the accuracy in the learned conditions for each completeness level separately. (D) Effect sizes (cohen’s d ± CI) of the data shown in (A), compared to 0 (no bias). Positive values indicate a pattern completion bias, negative values indicate a pattern separation bias.