Applying the Kessler Foundation Modified Story Memory Technique to Improve New Learning in Aging Populations

Cognitive decline is common with advancing age, most notably in learning and memory. While memory failures occur within the context of advancing age, increased difficulties have been noted in a subgroup of older individuals meeting the criteria for mild cognitive impairment (MCI). MCI lies between healthy aging and dementia, showing cognitive decline not severe enough to impair instrumental activities of daily living [1, 2]. Prevalence rates for MCI range from 7 to 25% over age 60 [1, 3‒5], with risk increasing with age [6]. Roughly 50% of those with MCI progress to Alzheimer’s disease over 4–5 years [7]. Most others continue to show subtle deficits but do not progress to dementia. Although, by definition, individuals with MCI and healthy older adults retain independence in instrumental activities of daily livings, the cognitive decline observed has been shown to negatively impact everyday functioning [8] and quality of life [9‒11], particularly in complex areas of daily life, such as scheduling/attending appointments and financial management [9]. Improvement in learning and memory would thus improve the everyday functioning and quality of life of healthy older adults as well as those with MCI.

While the identification of pharmacological treatments for cognition is a major emphasis of ongoing research, little benefit has been observed in aging populations [12‒14]. Non-pharmacological treatments such as exercise, life-style changes and cognitive training (CT) are often utilized and encouraged across populations to slow the progression of cognitive decline over time [15]. Side effects are few and participation may take place at home or in the community, increasing social engagement and social participation. CT has been advocated as a tool for maintaining and improving cognition in older adults [16] and has been tested in healthy adults, MCI, and Alzheimer’s disease, with positive results [17]. CT is thus a potentially powerful tool for optimizing cognition in healthy older individuals [18] as well as persons with MCI [19]. Individuals with MCI present with mild to moderate new learning and memory deficits, with retention of many other non-learning and memory cognitive capacities, making them ideal candidates for CT [19].

The Kessler Foundation modified Story Memory Technique (KF-mSMT^®) is a strategy-based CT protocol that has demonstrated efficacy to improve new learning in memory in several neurological conditions. The KF-mSMT^® is a 10-session strategy-based, manualized treatment protocol that teaches the strategies of imagery and context to improve learning [20‒22]. Previous research with the KF-mSMT^® has consistently highlighted its efficacy across populations, with post-treatment improvement on objective learning tasks [20‒22], self-report of improvements in daily life [20‒22], changes in patterns of cerebral activation [23, 24] and increased functional connectivity [25]. The treatment effect has been demonstrated to be maintained 6-months after treatment [20, 22]. These convincing data provide Class I evidence supporting the efficacy of the KF-mSMT^® for improving learning in TBI and MS [20, 22] and yields promise for improving learning in other populations, such as aging and MCI.

The current study tested the efficacy of the KF-mSMT^® in older individuals, both those meeting criteria for MCI and those deemed cognitively healthy. We hypothesized that individuals who completed treatment would show improvement on measures of learning and memory, while those assigned to the placebo control group would not.

Recruitment ran from March 1, 2018, to July 20, 2020. Participants were recruited through the use of an IRB-approved advertisement distributed in the community, including senior centers, community centers and events, libraries, and places of worship. In addition, participants who had previously participated in studies at Kessler Foundation were invited to participate in the current study as their information is stored in our participant database (with their permission, as obtained during the consenting procedure at their previous visit). The Subject Information Management System (SIMS) database encrypts PHI to keep the information private and protected. Thirty individuals over age 65 were randomized to the treatment group (TX: n = 17) or placebo control group (pCTL: n = 13). Four participants dropped out of the study prior to immediate follow-up, resulting in a final n of 26 (TX = 14; pCTL = 12). One of these participants never started treatment due to scheduling issues, one was excluded due to hearing and vision impairments and the other 2 dropped out mid-intervention due to loss of interest. Attrition thus fell at 13.33%. See Figure 1 for participant flow. There were no statistically significant differences between groups in regard to demographic characteristics, pretreatment cognition, and emotional symptomatology (Table 1). Inclusion criteria were (1) age 65+; (2) intact verbal comprehension as assessed with the Token Test; (3) no neurologic history or current diagnosis; (4) no history of major depressive disorder, schizophrenia, or bipolar disorder; (5) no history of diagnosed substance use disorder; (6) at least 20/70 vision in the worst eye on the Rosenbaum Eye Test.

This RCT used a 5-week, double-blind, parallel-group design. Before enrollment, all potential participants were screened for inclusion criteria.

Using a computerized random number generator, 40 potential participants were randomized to the treatment or placebo control group before beginning data collection. Data collection was discontinued at the end of the funding period, at which point 30 participants were randomized; thus, randomization was not balanced at trial completion. Treatment allocation was concealed. The individual responsible for group assignment was not otherwise involved in data collection, and group assignment was verified by a second individual via duplicate copy of the randomization. After randomization, participants completed baseline testing, consisting of a neuropsychological assessment and questionnaires assessing everyday cognition. Follow-up assessment was conducted within 1 week of completing CT using the same procedures. All memory measures utilized alternate forms.

The same research assistant (RA) conducted baseline and immediate follow-up evaluations, blinded to group membership. Masking was preserved via the following procedures: (1) the RAs conducting the treatment and assessments were different, and (2) the RAs did not communicate directly about participants. Study participants were also blinded to group assignment. Participants consented to participate in a study examining the impact of memory exercises in which they had a 50/50 chance of being in the treatment group. All participants completed a post-study questionnaire; 50% of participants correctly identified group assignment, while 50% of participants responded incorrectly or refused to guess, with no difference in response by group assignment.

The treatment group completed the KF-mSMT^®. Sessions occurred twice per week for 5 weeks, lasting 45–60 min each. The KF-mSMT^® group received the KF-mSMT^® training 2×/week for 5 weeks (10 sessions). In sessions 1–4 of the KF-mSMT^®, the participant learns to form mental images of the to-be-learned information through the presentation of a story with 20 to-be-remembered words displayed. In sessions 5–8 of the KF-mSMT^®, the participant learns to utilize context to facilitate learning. A list of to-be-remembered words is provided. The participant is instructed to create a story (context) with the words and then apply the imagery learned in sessions 1–4. Sessions 9–10 focus on applying the KF-mSMT^® to real-world settings. The therapist follows a training manual.

The placebo control group (pCTL) meets with the therapist at the same frequency as the KF-mSMT^® group, engaging in non-training tasks to control for professional contact. In sessions 1–4, pCTL group reads the same stories as the KF-mSMT^® group, but instead of learning context and imagery, participants in the pCTL answer questions about the story content. In sessions 5–10, participants in the pCTL group read stories in which the target words from treatment are embedded and answers questions about the stories. The pCTL condition is matched to the KF-mSMT^® condition in the duration of contact with the examiner and the medium of presentation. The only difference is that the pCTL group is not exposed to the critical ingredients of context and imagery that form the core of the KF-mSMT^®. Previous work on the KF-mSMT^® has shown the pCTL condition to be well-matched to the KF-mSMT^®condition; participants are not able to determine what group they were assigned to post-treatment, with the probability of participants guessing their groups falling under 50% in our previous RCTs [20, 22].

The primary outcome was performance on the California Verbal Learning Test (2nd edition; CVLT-II [26]) from before to after treatment, an objective cognitive performance test of learning and memory. The CVLT-II consists of a list of 16 words from 4 semantic categories presented orally over 5 trials and includes a 20-min delayed recall trial as well as a recognition trial. The CVLT-II has two alternate forms to minimize carryover between repeat testing sessions. No significant differences have been noted between the forms. The CVLT-I [27] was used as the third form. There are data indicating the CVLT-I and CVLT-II to be highly correlated. Research has shown both the CVLT-I and the CVLT-II to have good reliability and validity. The CVLT-II has shown high internal consistency reliability (r = 0.94) and split half reliability (r = 0.83). Test-retest reliability was also high for the CVLT-II (r = 0.82). The CVLT-II demonstrates good construct validity. Research has shown the CVLT-II to have high correlations with its predecessor, the CVLT-I (0.76 for total recall; 0.78 for long delay free recall). In addition, the CVLT-II has been noted to have a similar factor solution to its predecessor.

In addition to examining the primary outcome via ANCOVA, we examined the percentage of participants showing improvement following treatment, as is customary in the pharmacological literature examining cognitive changes with treatment [28]. In this analysis, benefit was defined as a 10% or greater improvement on the CVLT-II total learning from trials 1–5 from baseline to immediate follow-up.

Secondary outcomes included a subjective report of everyday cognitive functioning, using the Memory Functioning Questionnaire. This questionnaire was designed to quantify self-reported memory complaints. It consists of 64 items, presented in seven sections, each rated on a 7-point scale, which refers to frequency, quality, or seriousness of the memory complaint. The MFQ has been shown to demonstrate high test-retest reliability as indicated by a covariance analysis of longitudinal data, which was non-significant [29]. Internal consistency estimates are also high for the four factor scores resulting from the MFQ (Cronbach’s alpha = 0.94, 0.94, 0.89, 0.83). This indicates that the measure is highly reliable. Importantly, this factor structure is invariant across age, over different samples, and longitudinally [29].

Depression and anxiety were measured by the Chicago Multidimensional Depression Inventory (CMDI) and the State Trait Anxiety Inventory (STAI), respectively. All primary and secondary outcomes were identified prior to beginning data collection.

Statistical analyses were completed with intention to treat analysis with the last observation carried forward to address missing data. Differences between the two groups at baseline were assessed with independent sample t tests. Data were analyzed via analysis of covariance (ANCOVA), with follow-up performance as the outcome, group assignment as the predictor, and baseline performance as the covariate. To examine the percentages of participants that benefited from treatment, a chi-square (X²) was conducted. Analyses were performed with SPSS version 18 (SPSS Inc., Chicago, IL, USA). No interim analyses were performed.

After participants completed the baseline cognitive evaluation, performance was compared to normative data independently by two licensed psychologists to determine if they met criteria for diagnosis of MCI. For data analysis purposes only, participants that were determined to meet DSM-V criteria for MCI by both psychologists were classified as such (n = 19; TX = 10; pCTL = 9) while participants that did not meet DSM-V criteria for MCI were classified as healthy aging (n = 11; TX = 7; pCTL = 4). In the case of disagreement, the 2 psychologists discussed the profile and reached an agreement. This allowed the examination of the impact of cognitive impairment on treatment efficacy via ANCOVA, with follow-up performance as the outcome, group assignment as the predictor, and 2 covariates (baseline performance and cognitive classification).

An institutional review board approved all study procedures. All participants provided written informed consent. The clinical trial is registered with clinicaltrials.gov (protocol ID: NCT03370224).

Neuropsychological performance of the treatment and control groups at each baseline can be found in Table 2.

After controlling for variance in the CVLT-II total learning at baseline, the treatment group demonstrated a trend toward significance on total learning across the 5 learning trials at immediate follow-up compared with the placebo control group (F(1, 30) = 3.59, p = 0.069; Fig. 2), demonstrating a large effect size (partial eta squared = 0.12). The treatment group was able to learn more words than the placebo control group consistently across the 5 trials after treatment, while the 2 groups did not differ prior to treatment.

With benefit defined as a 10% or greater improvement on the CVLT-II total learning from trials 1–5 from baseline to immediate follow-up, results indicated that 65% of patients in the treatment group showed improvement on their CVLT-II total learning across trials 1–5 from baseline to immediate follow-up, as compared with only 31% of the control group (x²(1) = 3.4, p = 0.06).

After controlling for baseline, the treatment group reported significant improvements on the MFQ post-treatment as compared with the placebo control group, noted on frequency of forgetting when reading a novel (F(1, 27) = 4.18, p = 0.05, partial eta squared = 0.15, large effect). A trend toward significant improvement was noted on frequency of forgetting when remembering events (F(1, 27) = 3.7, p = 0.066; partial eta squared = 0.13; large effect). Other comparisons on the MFQ showed medium effect sizes, but were not statistically significant, namely, specific frequency of forgetting (F(1, 27) = 2.05, p = 0.17; partial eta squared = 0.08, medium effect) and the total MFQ score (F(1, 27) = 1.82, p = 0.19, partial eta squared = 0.07, medium effect).

No significant changes in depression or anxiety were noted.

To examine the role of level of cognitive impairment in treatment efficacy, ANCOVA was conducted with participants in the treatment group only. The covariate was CVLT-II performance at baseline, while classification of impairment level (MCI vs. HA) served as the fixed factor (IV). After controlling for variance in the CVLT-II total learning at baseline, the classification of healthy aging or MCI was not statistically significant. However, we did note a large effect size for this covariate, indicating that with a larger sample, degree of cognitive impairment could be expected to serve as a significant covariate (F(1, 17) = 2.16, p = 0.16, partial eta squared = 0.13, large effect; Fig. 3). The group classified as healthy aging, due to at least average scores on baseline cognitive testing, showed greater benefit from treatment than the group classified as MCI. No significant difference between the healthy aging and MCI groups were noted on the MFQ, depression, or anxiety from before to after treatment.

Results of this double-blind, placebo-controlled, randomized control trial support the utilization of the KF-mSMT^® in improving learning in the aging population. A significant impact of the KF-mSMT^® was demonstrated on an objective measure of new learning and memory, the gold standard for the evaluation of the efficacy of cognitive rehabilitation, with a large effect size noted. Greater than 10% improvement was documented in 65% of those treated with the KF-mSMT^®; this is compared with only 31% of the control group, the latter likely attributable to practice effects, as observed in previous work on the KF-mSMT^® [20, 22]. These results provide preliminary evidence indicating that the KF-mSMT^® is effective when evaluated via traditional criteria used to evaluate pharmaceuticals for cognitive improvement [28].

Beyond the laboratory, it is also important to note that patients that were treated with the KF-mSMT^® themselves reported improved memory functioning in their daily life, in contrast to the placebo control group, who reported declining memory abilities in daily life. While existing studies support memory strategy training, with gains remaining over several months [30, 31], published data regarding the generalization of this improvement to everyday life remain limited [32, 33]. Generalization to daily life is a major emphasis in aging and dementia research, as data regarding the impact of cognitive rehabilitation on everyday outcomes in this group are lacking [16, 34]. The current trial begins to address this void in the literature.

The current study provides preliminary proof of concept that the KF-mSMT^® addresses new learning challenges that aging individuals face in their daily lives and may serve to reduce the burden of these challenges. There were no significant differences between participants classified as having cognitive abilities consistent with healthy aging or MCI, suggesting that both groups can benefit from utilization of the KF-mSMT^®. However, a large effect size was noted in this analysis indicating that with a larger sample a significant difference may in fact be noted between these groups. That is, comparison of mean performance between the groups indicates that, while all participants showed benefit from the KF-mSMT^®, the healthy participants showed greater benefit than individuals with cognitive impairment consistent with MCI. Given the small sample size, it is important that these results be replicated with larger samples as they have significant implications for the management of memory abilities as individuals age. That is, consistent with much of the recent literature [35, 36], cognitively enhancing interventions have a greater impact prior to the onset of memory deficits. This indicates that it may be beneficial for aging individuals to strengthen their memory abilities while cognitively healthy in an effort to maximize their memory abilities with advancing age. Longitudinal research is necessary to examine the long-term impact of the KF-mSMT^® addressing the specific question as to the ability of such enhancement to stave off the memory deficits characteristic of MCI. Indeed, one large trial in aging demonstrated that cognitive enhancement delayed the onset of dementia up to 10 years following treatment [37]. More longitudinal work such as this is necessary to demonstrate the longer-term benefit of such treatment.

While results are encouraging, several limitations of the current study highlight the need for future research applying the KF-mSMT^® to aging populations. The first is sample size. With only 17 participants in the treatment group and 13 in the control group, statistical analyses were substantially limited by the available power. Effect size was used to interpret the potential impact of a larger sample and clearly supports a larger RCT to fully evaluate the KF-mSMT^® in aging. Those classified as healthy aging versus MCI was similarly small, highlighting the need for future research in this area. In addition, the study utilized tests common to cognitive rehabilitation research but less commonly utilized in aging and dementia populations. Replication of these results utilizing measures known to be sensitive to the effects of MCI and aging on cognition and daily life function is needed. Additionally, longitudinal research is needed to determine whether the effects observed in this study persist beyond the immediate assessment. Finally, the inclusion of neuroimaging pre and post-treatment would allow the delineation of the mechanisms underlying the efficacy of the KF-mSMT^® in older adults, as done in other populations [23‒25].

Results of the current study present preliminary proof of concept for the application of the KF-mSMT^® to improve new learning and memory in the aging population and provides preliminary evidence supporting the need for a larger clinical trial. That is, we observed improved new learning and memory performance following completion of the KF-mSMT^® in both individuals who meet criteria for MCI and those who do not. This indicates that the KF-mSMT^® could potentially be useful for improving new learning and memory abilities in this population.

This study protocol was reviewed and approved the Kessler Foundation Institutional Review Board, Approval No. R-990-17. All human subjects provided written informed consent.

The authors have no conflicts of interest to report. KF-mSMT^® is a registered trademark owned by the Kessler Foundation. All authors were Kessler Foundation employees when the study was conducted.

This research was supported by the New Jersey Health Foundation Pilot Grant and Kessler Foundation.

N.D.C.: conceptualization of study, grant funding acquisition, data collection, data interpretation, and initial draft of manuscript. E.W. and S.L.C.: data analysis, data interpretation, and critical review of manuscript. J.L., D.M., and N.B.M.: data collection and critical review of manuscript. J.D.: conceptualization, grant funding acquisition, and critical review of manuscript. All authors granted final approval of the version to be published and agree to be accountable for all aspects of the work.

The data that support the findings of this study are not publicly available due to content of information that could compromise the privacy of research participants but are available from the corresponding author (N.D.C.) upon reasonable request.