Associative-memory Deficit As A Function Of Age And Stimuli Serial Position Part 1

Abstract

Studies have shown associative-memory decline in aging. While the literature is inconclusive regarding the source of the deficit, some researchers argue that it is caused by impaired encoding and maintenance processes in working-memory (WM). 

Successful retrieval of a stimulus depends on its sequential presentation in the learning list: stimuli at the beginning or the end of the learning list benefit from higher retrieval probability. These effects are known as "primacy" and "recency" effects, respectively. In the case of the primacy-effect, stimuli at early list positions benefit from extensive rehearsal that results in enhanced consolidation and trace in long-term memory (LTM). In the case of the recency-effect, target stimuli at later serial positions are still maintained in WM and can therefore be effortlessly retrieved. 

Considering these effects could shed light on the involvement of WM in associative-binding. Both behavioral and neuroimaging researchers have studied associative-decline in aging. However, no work has explicitly tested age differences in memory for items versus associations as a function of stimuli serial position (SSP). 

In the current study, 22 younger and 22 older adults were recruited to participate in a study aimed to test the separate and joint effects of both SSP and aging on memory-recognition of items and associations. In the task used, retrieval was manipulated for SSP (beginning/middle/end of the list) and item/associations recognition modes. We hypothesized that greater associative-decline will be observed in older adults, specifically for recently presented material. 

The results showed that both groups presented a significant associative-deficit at the recency positions; this decrease was additive and did not correspond to the expected interaction effect. Further analysis showed that the source of associative-memory decline for stimuli at recency position in older adults resulted from an increase in false-alarm (FA) rates. These results support the involvement of WM-binding impairment in aging.

Introduction

Numerous studies have shown age-dependent episodic memory decline [1]. Several hypotheses have been suggested to explain older adults' poor memory performance: reduction in attentional resources or processing abilities [2, 3], reduction in processing speed [4] and failure of inhibitory processes [5]. Notwithstanding, not all memory processes and components are similarly affected by age [6, 7], rather greater memory decline is observed in older adults when trying to generate and retrieve links between units of information as opposed to the retrieval of single units of information [8–11]. 

Multifaceted/complex episodes are based on multiple kinds of (eventually related) information sources. Remembering an episode requires remembering the individual pieces of information (i.e., "items") as well as their associations with each other and other specific or contextual information (i.e., "associations") [12, 13]. Studies conducted with different age groups have shown that older adults recalled significantly fewer targets and links between targets than did younger adults [14–16]. 

These empirical research findings in older adults brought Naveh-Benjamin [17] to conclude that with age, older adults mainly fail to encode and retrieve links between units of information (i.e., associative-memory). 

The associative-decline was formulated as the difference between younger and older adults in memory recognition for single units of information (i.e., items) versus associations between stimuli and was named the Associative-Deficit Hypothesis (ADH). 

This age-related associative-deficit is extensively supported by behavioral data [18–23] and applies to memory for paired-stimuli as well as to memory for source, context, temporal order and location, all of which require binding processes [10].

The literature is inconclusive regarding the source of the disproportionate deficit in associative versus single (i.e., items) units of information recognition. Some researchers argue that it is caused by impaired encoding and maintenance processes in working memory (WM). 

Cowan, Naveh-Benjamin, Kilb, & Saults [24] used visual objects and their spatial location in WM to assess for item change and binding deficits in younger and older adult participants. The results of the experiment showed a strong bias toward the detection of changes, especially binding changes. Chen and Naveh-Benjamin [25] used a continuous recognition paradigm and replicated an associative WM decline in three experiments. Brockmole, Parra, Della Sala, & Logie [26] used a change detection task to assess associative-memory decline and reported that older adults store proportionally fewer bound representations than individual features compared to younger adults. 

Lastly, Hara & Naveh-Benjamin [27] simulated long-term memory (LTM)- dependent associative-memory deficit (as seen in older adults) in young adults by manipulating WM resources, thus highlighting the involvement of WM in successful associative recognition. 

In contrast to these findings, other researchers did not find any associativedeficit in WM, which conflicts with the WM explanation. Bopp & Verhaeghen [28] addressed the WM associative-deficit question using a repetition-detection task that can differentiate memory for content from memory for context. 

In three experiments the authors did not find a specific age-related deficit for context in WM. 

Parra, Abrahams, Logie & Sala [29] addressed visual short-term memory using color-shape conjunctions to test associative-binding deficits in WM and also concluded that binding in WM is not age-dependent.

Taking into account the original serial position curve effects could shed light on the involvement (or lack of involvement) of WM in associative-binding. It is well known that the successful retrieval of a stimulus depends on its sequential presentation (beginning/middle/ end of the list) in the learning list [30, 31]. Stimuli at the beginning [32, 33] or the end [30, 34– 36] of the learning list benefit from higher retrieval probability compared to stimuli presented at intermediate positions. 

These effects are known as "primacy" and "recency" effects respectively. Neuroimaging studies have documented the involvement of the prefrontal cortex (PFC) [37–40] and structures in the medial temporal lobe (MTL), with a specific contribution of the hippocampus [41–44]. In the case of the primacy effect, extensive rehearsal for stimuli at early list positions results in better retrieval probability; thus, the primacy explanation links LTM formation and WM processes [45–47]. 

Neuroimaging studies applying functional magnetic resonance imaging (fMRI) support differential involvement of MTL structures in this process by showing that WM maintenance facilitates the encoding of a stimulus into LTM by activating rehearsal processes in the hippocampus (for associations between stimuli [48]) or in the parahippocampal cortex (for single units of information [49]). In the case of the recency effect, target stimuli at late serial positions (i.e., at the end of the learning list) also benefit from higher retrieval probability because they are still maintained in WM and therefore can be effortlessly retrieved, compared to stimuli at intermediate positions [30, 46, 50–52].

Reports of studies that tested the effect of the serial position of a stimulus on memory in older adult participants have indicated that while the primacy effect was absent, the recency effect remained intact [53, 54]; these studies however did not account for associative-memory. Lately, we simulated an associative-memory deficit (as seen in older adults) in young adult participants by controlling stimuli serial position (SSP) and presentation duration. This resulted in greater memory decline for associative material presented at the end of the learning list compared to single-unit targets (i.e., items) at similar positions [55]. These results highlight the different benefits single units of information have from a late sequential position during learning as compared to paired units of information. The results raise a question regarding the nature of both intact as well as impaired associative-binding in WM.

Because the recency effect reflects retrieval from WM, it is expected that impaired WM will contribute to the associative-deficit in older age; in that case, the age-related associative-deficit should be greatest for the recency portion of the list compared to the beginning and middle portions of the list. If the age-related associative-deficit is similar in all parts of the list, regardless of the serial position of the studied information, then this deficit cannot be attributed solely to impaired WM. Whereas both behavioral and neuroimaging researchers have studied associative-decline in aging, currently no work has explicitly compared age differences in memory for items and associations as a function of SSP. 

In the current study, 22 younger and 22 older adults were recruited to participate in a study aimed to test the separate and joint effect of both SSP and aging on memory recognition for items versus associations. We hypothesized that greater associative-decline (compared to the expected decline in memory for items with similar serial location) will be observed in older adults, specifically for recently presented material (i.e., stimuli that was presented at the end of the learning list).

Materials and methods

Participants

The number of participants was calculated based on our last study which was conducted with young adult participants and had tested similar effects [55]. Calculation of the estimated effect and sample size was conducted using MedCalc software and considered both type-1 (α. 0.05) and type-2 (β, 0.1) errors as well as the estimated difference between means (0.26, based on previous research) and corresponding SD (0.18, 0.29 based on previous research) for each group. Using G�power software we calculated the estimated minimum number of participants for the total sample according to the estimated effect size (η2 p = 0.25) of the highest hypothesized interaction (F-test, ANOVA: repeated measures, within-between interaction). 

This calculation resulted in ~20 participants for each group and served as the basis for the estimated number of participants that were invited to take part in the current study. Based on this calculation, 46 participants were invited to participate in the current study, of which 22 were younger (M(years) = 24.90±2.12 SD, 9 women) and 24 were older (M(years) = 73.61±8.10 SD, 14 women) adults. 

All reports were given by the participants via self-report. All participants reported normal and intact everyday functioning with no disabilities and/or psychiatric disorders. Participants in the younger adult group were Achva Academic College students that were rewarded for their participation with course credit, an acceptable procedure in a first-year introductory psychology academic course. The older adults group consisted of participants recruited from the local community and senior citizens' home. Participants and their caregivers were asked about everyday functioning. 

They were asked to describe and report any cognitive/physical disabilities. Exclusion criteria included past/current psychiatric or neurological disorders, current sensory/motor disorders and/or a formal diagnosis of learning disabilities. Participants that met one or more exclusion criteria in their description were excluded from the study. Two (older) participants were excluded from the study and were not included in the analysis after their caregivers reported that they were diagnosed with mild cognitive impairment (MCI). 

Finally, 22 participants for each group were included in the analysis. The study was approved by the local institutional review board of Achva Academic College. All participants gave their written informed consent for participation in the study.

Experimental design and hypotheses

Three independent variables were used in this study: SSP (beginning, middle and end of the learning list; a within-subject variable); X test (item versus associative recognition; a withinsubject variable); and X age (younger versus older adults; a between-subject variable). The dependent variable was memory accuracy, calculated as "hit minus false-alarm (FA)" rate for each participant in each experimental condition. 

In addition, and to specifically address a cumulative associative-decline, we computed associative-deficit index (ADI) reflecting the difference between item recognition and associative recognition performance. The ADI was calculated by subtracting the proportion of Hits minus the proportion of false alarms in associative recognition trails, from the proportion of Hits minus the proportion of false alarms in item recognition trails (proportion of Hits-FAitem minus proportion of Hits-FAassociation). 

Higher ADI scores reflect greater associative-deficit. While differences in item and associative recognition were evident for all participants regardless of age, our main hypothesis was that the greatest difference between item and associative recognition (i.e., associative-deficit, as measured via ADI scores) would be evident for older adults and for stimuli located at the end of the learning list (compared to material located at the beginning/middle of the learning list).

Memory task

24 separate lists of 12 pairs of words were built from a pool that contained 576 words that were unrelated visually, semantically and auditorily. The lists consisted of high-frequency Hebrew common nouns (based on the Hebrew norms [56]). After each study list, participants immediately performed an item or an associative test for stimuli located in differential list positions; beginning (4 first pairs of the learning list, to assess the primacy effect), middle (4 middle pairs of the learning list) and end (4 last pairs of the learning list, to assess the recency effect) of the learning list. 

Stimuli were presented visually and displayed centrally on a 15" computer screen one at a time. The four replications for each of the six conditions (3(stimuli serial position) X 2(test)) were presented randomly across participants and each stimulus was used only for one of the tests (i.e., each stimulus was only used once in the experiment). Fig 1 describes the experimental paradigm.

Item recognition test

Participants were presented with 4 words taken from only one list position (beginning/middle/end). Of the 4 words, 2 words were targets, i.e., words that had appeared in the learning list, and 2 were distracters, i.e., newly introduced words. Participants were informed that the list include targets and distracters, and were instructed to respond as quickly and accurately as possible to each stimulus with a designated "yes" key for targets and a "no" response key for distracters.

Associative recognition test

Participants were presented with 4 pairs of words taken from only one list position (beginning/middle/end). Of the pairs, 2 pairs were targets, i.e., word pairs that had appeared in the learning list and 2 were distracters, i.e. rearranged pairs. Distracters were items that had appeared in the learning list but were now recombined into novel (distracter) pairs. Recombined pairs were created from similar list positions and were not mixed across the list. Participants were again informed that the list included targets and novel pairs, and were instructed to respond as quickly and accurately as possible to each pair with the same keys as in the item recognition test.

Procedure

Younger participants were tested individually in designated rooms at Achva Academic College and older participants in the senior citizens' home. Participants were given instructions and also clarifications as needed before the start of the first learning list. Participants were informed that they were about to view 24 learning lists and that each list would be followed by either an item or an association test. 

Participants were asked to learn and remember both the individual items as well as the pairs presented in the learning list for the upcoming tests. In each test, participants viewed 12 word pairs on the computer monitor, one at a time, at a rate of 2 seconds per pair. All tests in the current experiment were self-paced; that is, the tested material appeared on the screen for 1.5 seconds. 

A response during this time range (1.5 seconds) caused the appearance of the next stimulus. If no response was recorded within 1.5 seconds, the stimulus disappeared from the screen, and the next stimulus appeared only after a response had been recorded. In all test steps, only the response to a stimulus caused the appearance of the next one. In each memory test type (item recognition or associative recognition) participants were tested for only one of the three stimuli locations: four words located at the beginning of the learning list (for testing the primacy effect), four words located at the middle of the learning list and an additional four words located at the end of the learning list (for testing the recency effect). 

Each stimuli list location was tested individually with a different test. Participants were blind to stimulus location before the actual test phase, i.e., they were presented with an item or an association test, but were not introduced to the stimulus location from the learning list until the beginning of the test.

For more information:1950477648nn@gmail.com