Effects of healthy ageing on precision and binding of object location in visual short term memory

Abstract

Visual short term memory (STM) declines as people get older, but the nature of this deterioration is not well understood. We tested 139 healthy subjects (19-83 years) who were first required to identify a previously seen object and then report its location using a touchscreen. Results demonstrated an age-related decline in both object identification and localization. Deterioration in localization performance was apparent even when only 1 item had to be remembered, worsening disproportionately with increasing memory load. Thus, age-dependent memory degradation cannot be explained simply by a decrease in the number of items that can be held in visual STM but rather by the precision with which they are recalled. More important, there was no evidence for a significant decrease in object-location binding with increasing age. Thus, although precision for object identity and location declines with age, the ability to associate object identity to its location seems to remain unimpaired. As it has been reported that binding deficits in STM might be the first cognitive signs of early Alzheimer's disease (AD), the finding that object-location binding processes are relatively intact with normal aging supports the possible suitability of using misbinding as an index measures for probing early diagnosis of AD.

Figure 1

Procedure of a three-item trial: After the initial presentation of three items and after a delay (1 or 4 s) subjects had to recognize and touch the target that had appeared in the memory array of this trial and then drag it to its original location from the memory array. See the online article for the color version of this figure.

Figure 2

Identification accuracy. (a) Schematic representation of the two-alternative forced-choice identification task. (b) Proportion of trials where the target was identified correctly shown for each of the four different age groups. The x-axis represents the delay duration. Black lines denote one item conditions; red (gray) ones indicate three item conditions. See the online article for the color version of this figure.

Figure 3

Localization performance. (a) Schematic representations of the different measures (upper figure: error in a three-items trial; below: NIC to the closest object. Green (dark gray circles filled with white) circle represent the target’s original location and red (bright gray) circles the nontarget locations. (b) Localization error for the four different age groups: The x-axis represents the maintenance delay. Red (gray) are three items conditions; black are one item conditions; blue (dashed) lines are nearest item control (NIC) measures. See the online article for the color version of this figure.

Figure 4

Precision of recall with increasing age. (a) Linear regression between age and localization error when one item has to be remembered, (b) when three items have to be remembered, and (c) for nearest item control (NIC) values. There is an age-related increase of localization error when only one item and three items are remembered, as well as an increase of NIC values (localization error controlled for object-location swapping). The 95% confidence intervals are marked in gray shade. See the online article for the color version of this figure.

Figure 5

(a) Linear regression between age and difference in localization error between three items and one item to-be-remembered. (b) Linear regression between age and difference in localization error between NIC and one item to-be-remembered. The 95% confidence intervals are marked in gray shade. See the online article for the color version of this figure.

Figure 6

Swap errors for different age groups. (a) Schematic representation of a swap error (green [dark gray filled with white] circle represents the target’s original location and red [light gray] circles the nontarget locations). (b) Proportion of trials where the target has been located to a nontarget location for the four different age groups. The x-axis represents the maintenance delay. See the online article for the color version of this figure.

Figure 7

Swap errors and aging. (a) Linear regression between age and overall proportion of trials where swap errors occurred. (b) Linear regression between age and corrected swap error rate. The 95% confidence intervals are marked in gray shade. See the online article for the color version of this figure.