Is Cognitive Training Worth It? Exploring Individuals’ Willingness to Engage in Cognitive Training

Abstract

We assessed how much time individuals would be willing to spend engaging in game-based cognitive training to gain prolonged functional independence. In Study 1 (N = 294), participants completed a survey with questions assessing how much time they would be willing to invest in daily cognitive training to extend their functional independence by certain amounts of time using a slider response that ranged from 0 to 100 minutes. Participants also completed surveys that measured self-perceived health and cognitive functioning, personality, and other demographic variables. Even for relatively small gains, participants reported being willing to dedicate an average of 11 minutes every day to cognitive training, with some participants willing to engage for significantly longer. The best predictor of willingness to invest time in training was belief in cognitive training efficacy, followed by openness to experience, and participants’ self-perceived cognitive deficit. Study 2 examined the same question in a sample of 120 older adults, this time allowing for open-ended responses. Participants reported being willing to invest significantly more time, ranging from more than 40 minutes every day to gain just one week of independence, to over 2.5 hours every day to gain an additional 3 years of independence. Again, perception of cognitive training efficacy was the strongest predictor of willingness to invest time. Results confirm that older adults are willing to invest significant amounts of time to gain independence later in life, and have implications for predicting the adoption of, and adherence to, potentially effective treatments for cognitive decline.

Introduction

In recent years there has been a rise in the popularity of “brain training” or cognitive training programs (e.g., Lumosity, CogniFit, Brain HQ).¹ These programs typically involve digital games or gamified neuropsychological tests designed to exercise fundamental cognitive abilities, with the aim of improving the performance of important everyday tasks (e.g., school or work performance, driving, remembering names and faces). These programs promise improved memory, increased attention span, protection against memory loss, and even increased IQ, and often target older adults who may be experiencing age-related cognitive decline. However, there is no clear consensus regarding the effectiveness of cognitive training, and this topic has engendered a great deal of controversy. In 2014, an international group of 70 scientists published a letter asserting that cognitive training does not provide a scientifically grounded way to prevent cognitive decline or improve cognitive functioning (“A Consensus on the Brain Training Industry from the Scientific Community,” 2014). That same year, 133 scientists and practitioners published a counter-statement highlighting the abundance of evidence for the benefits of cognitive training (“Cognitive Training Data,” 2014). Further, after reviewing evidence for cognitive outcome improvements put forward by this group, others not involved in either statement found much of the provided evidence inadequate (Simons et al., 2016). In general, this field of investigation is characterized by a lack of consensus.

Aside from the controversy regarding whether or not cognitive training works, a separate, but potentially equally important question is whether older adults are willing to engage in cognitive training over an extended period of time. Should effective training programs be discovered, they would have little impact unless older adults adopt these programs and adhere to them. As one example, consider the study conducted by Boot et al. (2013). Evidence at the time suggested that action video game play was associated with improved perceptual and cognitive abilities (e.g., Green & Bavelier, 2010), leading Boot and colleagues to develop an action video game intervention involving the popular game Mario Kart for older adults. Unfortunately, no cognitive benefits were observed as a result of the action game intervention. However, this failure may in part have been due to older adults failing to adopt and adhere to the intervention. Although participants were asked to play for 60 hours over the course of three months, they in fact only played 22 hours (M = 22 h, SD = 5; excluding 7 of 21 participants who dropped out of the study entirely). However, in another condition, participants who played a game that focused specifically on cognitive fitness (Nintendo’s Brain Age 2) adhered to the intervention and played close to the 60 hours requested (M = 56 h, SD = 6; excluding only 1 out of 20 participants who dropped out of this condition). Participants in this group reported the belief that training with this game was more likely to improve their performance of important everyday tasks such as driving.

What motivates someone to adopt and adhere to a cognitive training program? How much time are people willing to invest in training to obtain benefits? In some ways, the decision to adopt a cognitive training program is analogous to a temporal discounting problem. Temporal discounting problems often ask individuals whether they would be willing to give up a smaller monetary reward now in order to obtain a larger reward in the future (e.g., Green, Myerson, & McFadden, 1997). Depending on the circumstances, individuals may reject a larger reward in the future for a smaller immediate reward (e.g., if the difference in reward amount is small). That is, they discount the value of future rewards. Analogously, in the case of cognitive training, older adults may give up a certain amount of time now engaging in training to increase their independence later in life as perceptual and cognitive abilities begin to decline (see Toril, Reales & Ballesteros, 2014, for additional discussion). This is not a hypothetical proposition; delayed reward is consistent with evidence suggesting that cognitive training benefits might not be immediate, but instead might manifest years later (Tennstedt & Unverzagt, 2013). Understanding willingness to invest time in cognitive training will require understanding the factors that influence older adults’ evaluation of the tradeoff between time invested now and benefits obtained later.

What general factors might enter into this decision to cause someone to either adopt or not adopt cognitive training? Decision-making models related to health activities provide some insight. The Health Belief Model (HBM) proposes that health-related actions and adherence depend on: 1) sufficient motivation or health concern to make a health issue important to an individual, 2) perceived threat, or the perceived vulnerability to a serious health problem, and 3) the belief that the perceived threat would be reduced by following a health recommendation, and that the cost of doing so is subjectively acceptable (Rosenstock, Strecher & Becker, 1988). These three classes of factors have clear analogues in the decision to adopt and adhere to cognitive training. If health concern and perceived threat are low (a person doesn’t believe they are likely to experience cognitive decline, or that these declines can impact their independence) and they believe cognitive training programs are ineffective, it is unlikely that they will adopt and adhere to training. These ideas are similar to the Protection-Motivation Theory of Rogers (1983) that posits that threat appraisal (perceptions of vulnerability and severity of an illness) and coping appraisal (belief that a recommended response is effective, easy to perform, and involves few costs) shape engagement in health behaviors.

Because many cognitive training programs are technology-based (delivered by computer or mobile device), the Senior Technology Acceptance Model may be especially relevant (Chen & Chan, 2014). This model predicts that attitudes toward using a piece of gerontechnology (technology to benefit older adults) are influenced primarily by the perceived usefulness and perceived ease of use of the technology. Self-efficacy, health, physical functioning, and cognitive ability, among other factors, influence perceived usefulness and ease of use. Regarding perceived usefulness and ease of use of technology, Mitzner and colleagues (2016) found that the personality factor of “openness to experience” positively predicted both.

These models, if applicable to the domain of cognitive training, predict that older adults’ perceptions that cognitive decline is a threat to them will increase their likelihood of adopting cognitive training, especially if they already have negative perceptions of their own cognition. However, this is in contrast to a recent online survey conducted by AARP to better understand adults’ willingness to include more cognitively stimulating activities into their daily routines (Mehegan, Rainville, & Skufca, 2017). This survey found that adults who reported higher, rather than lower, cognitive functioning and well-being had a greater desire to do more to improve their cognitive health and engage in cognitively stimulating activities. It was notable, though, that only 28% of individuals 40 years of age or older reported engaging in computer-based cognitive training regularly. Important individual difference variables may relate to this decision. For example, older adults may be more likely to adopt training if they perceive it as effective. Previous research has found that older adults are generally optimistic regarding cognitive training, but that there are individual differences in the strength of this belief (Rabipour, Andringa, Boot, & Davidson, 2018).

To understand how much time older adults might be willing to engage in cognitive training, and to investigate the influence that older adults’ beliefs have on their adoption of training, two studies were conducted that explored older adults’ willingness to engage in daily training with game-based cognitive interventions, starting now, to obtain an extension of a certain amount of time of functional independence later in life. Based on models discussed previously and the existing literature, we predict that key individual difference variables might affect decisions regarding cognitive training. We predict the following:

1. Participants will be sensitive to the amount of gained functional independence. They will be willing to invest more time daily for larger gains later.

2. Increasing age will be associated with a willingness to spend more time engaging in cognitive training due to worries about age-related cognitive decline. These concerns may be enhanced by personal experience with a close friend or family member experiencing dementia.

3. Those with poorer self-perceived health and cognition will be more willing to spend time on cognitive training to address their perceived deficit.

4. Personality may play a role. Specifically, openness to experience will be associated with more willingness to participate in cognitive training.

In addition to these primary hypotheses, we also present an exploratory set of analyses related to another question: Are the investments individuals report being willing to make sensible from the perspective of time invested versus benefit gained? For the individual, it may not be desirable to invest more time training than the total amount of time gained later in life through prolonged independence (though, as will be discussed later, this is a complicated issue that depends on a number of other factors). These exploratory analyses project the total amount of time participants report being willing to invest in daily cognitive training over different time scales, and compare this potential time invested to hypothetical extensions of functional independence (from 1 week to 3 years) later in life.

STUDY 1 - MECHANICAL TURK

Amazon’s Mechanical Turk service was used to collect a large dataset to assess individuals’ willingness to engage in daily cognitive training in order to achieve certain levels of benefit, and to explore predictors of willingness to engage.

Method

Participants

Participation was restricted to Mechanical Turk (MTurk) Masters in the United States. MTurk Masters are workers Amazon has judged to provide high quality responses. Further, workers were required to have 95% of their previous MTurk Human Intelligence Tasks (HITs) approved. Of the 463 MTurk workers responding to posted HITs, 294 were included in the reported analyses. Reasons for exclusion included failed attention checks, repeated responses from the same IP address, and failure to answer one or more survey questions necessary to compute a survey score. Participants (129 male, 165 female) ranged in age from 20–73 years with a mean age of 40.37 years (SD = 11.9 years). Fifty-five (18.7%) participants in the sample were fifty-five years of age or older. The final sample was 88% white, 6% black, 4% Asian, and 2% American Indian/Alaska Native, Multi-Racial, or “other.” All participants were paid 50 cents.

Procedure

All measures and procedures were approved by the FSU Human Subjects Committee. The MTurk HIT was described as a survey about “brain training,” personality, and health, with an estimated completion time of 20 minutes. Two HITS were created, one in which all ages could participate, and one that was only shown to MTurk workers 55 years of age or older to increase the representation of older participants. HITs were reposted every few hours to increase their visibility. After participants provided informed consent they were instructed to complete a number of online surveys.

Measures

Six measures were administered, along with a few additional questions about experience with cognitive training, and experience with friends and family members with dementia. All surveys and questions were administered using the Qualtrics survey platform. Of primary interest was the Brain Training and Independency (BTI) measure. This measure is described first, followed by other administered measures.

Brain Training and Independence Survey.

This measure, developed by Boot and colleagues, consisted of 7 questions intended to gauge how much time individuals were willing to devote to daily cognitive training for varying amounts of prolonged functional independence later in life. Participants were asked to assume that cognitive training was effective. After a description of game-based cognitive training (“These products usually take the form of simple games that can be played on a computer or mobile device”), the purpose of the measure, and example scenarios, participants were presented questions in the form of:

• If effective brain games were discovered that allowed people to live independently for 1 week longer, I would be willing to play these games: ______ minutes every day, starting today.

• If effective brain games were discovered that allowed people to live independently for 3 years longer, I would be willing to play these games: ______ minutes every day, starting today.

Specified periods of independence ranged from 1 week to 3 years (1 week, 1 month, 2 months, 6 months, 1 year, 2 years, 3 years). Participants responded by adjusting a slider that ranged from 0 to 100 minutes. We did not anticipate that participants would be willing to spend more than 100 minutes daily on brain training. The full measure can be found in Appendix A.

Background Information Questionnaire.

This measure included 14 questions that assessed sex, age, occupation, and other demographic variables (adapted from a survey developed by the Center for Research and Education on Aging and Technology Enhancement).

The General Self-Efficacy Scale.

To assess general self-efficacy, participants rated the veracity of 10 statements (e.g., “I can usually handle whatever comes my way,” Jerusalem & Schwarzer, 1995). End anchors were “Not at all True = 1”, and “Exactly True = 4.” Ratings were summed across questions.

Health Questionnaire.

A subset of 6 questions were administered from the Short Form Health Survey (SF-36, Ware & Sherbourne, 1992). Initially, the plan was to use a composite measure of all six questions, but some participants failed to answer one or more questions. Thus, only one question asking participants to rate their general health status was included in analyses. For participants who answered all heath questions, this composite health measure (extracted from factor analysis) was highly correlated with the single general health question (r = .85). Analyses did not meaningfully change when using either measure, thus we report analyses that preserve the greatest number of participants entering into the analyses.

Perceived Deficits Questionnaire.

This measure consisted of 20 questions that asked participants to rate their cognition, and was adapted from a scale developed by the National Multiple Sclerosis Society (Sullivan, 1990). Language specific to Multiple Sclerosis was removed. Scoring involved summing the response to all items, with higher scores corresponding to greater deficits.

Ten-Item Personality Inventory (TIPI).

Participants rated their agreement with ten statements describing their personality (Gosling, Rentfrow & Swann, 2003). Two questions each assessed Agreeableness, Conscientiousness, Emotional Stability, Extraversion, and Openness to Experience, and scores for each personality construct were obtained by averaging the two questions (higher scores indicating higher levels of that personality construct).

Other Questions.

Participants were asked if they had previous experience using cognitive training, if anyone close to them had experienced Alzheimer’s disease or other forms of dementia, and how effective they believed cognitive training to be.

Results

First, we examined the amount of time participants were willing to engage in game-based cognitive training as a function of potential gains in independence. As can be seen in Figure 1, there was a strong positive relationship between hypothetical independence gains and the amount of time participants were willing to invest. On average, participants were willing to devote between about 11 minutes (for 1 week of independence) and 42 minutes (for 3 years of independence) each day to cognitive training, though individual participants were willing to invest much more time (up to the maximum of 100 minutes).

Figure 1:

The mean number of minutes participants were willing to devote to cognitive training as a function of gains in independence for Study 1. Error bars = +/− 1 SEM.

Next, we explored the best predictors of willingness to engage in cognitive training. A regression analysis was conducted with total amount of cognitive training time (minutes summed across 7 questions) as the criterion variable, and then age, sex (−1 female, 1 male), self-efficacy, self-reported health, perceived cognitive deficit, the five factors from the TIPI, self-reported experiences with individuals with dementia (−1 No, 1 Yes), and previous cognitive training history as predictor variables (−1 No, 1 Yes, see Table 1). This model accounted for 21.6 percent of the variance in cognitive training time (F(13, 280) = 5.95, p < .001). The strongest predictor of reported cognitive training time investment was the perceived effectiveness of cognitive training, followed by openness to experience, followed by perceived cognitive deficit. The more participants believed in the efficacy of training, the more they were open to experience, and the higher their perceived cognitive deficit, the more willing they were to devote time to training. Note that an assumption of regression analysis is that model residuals are normally distributed. A Q-Q plot indicated violation of this assumption. Square root transformation of nondichotomous data greatly reduced this problem, and results did not meaningfully change (see OSF page for this supplementary analysis, https://osf.io/25czj/?view_only=100c1c46f2e04142a0e1a2efc5da41ba).

Table 1:

Regression predicting amount of time participants were willing to engage in daily cognitive training for Study 1.

Coefficients
Model	Unstandardized	SE	Standardized	t	p
1 (Intercept)	−23.769	102.392		−0.232	0.817
Age	1.056	0.751	0.081	1.406	0.161
Sex	−5.614	9.375	−0.036	−0.599	0.550
CTEffect	70.300	10.588	0.373	6.640	< .001
Health	−14.299	10.706	−0.084	−1.336	0.183
CogDef	1.683	0.808	0.130	2.084	0.038
Self-Efficacy	−2.198	2.080	−0.078	−1.056	0.292
Agreeableness	−2.573	7.578	−0.022	−0.339	0.734
Conscientiousness	8.429	7.830	0.069	1.077	0.283
Emotional Stability	−2.939	7.712	−0.029	−0.381	0.703
Extraversion	−0.318	5.177	−0.004	−0.061	0.951
Openness to Experience	15.258	6.409	0.143	2.381	0.018
Dementia Experience	1.712	8.670	0.011	0.198	0.844
Used Cognitive Training	−10.965	10.504	−0.057	−1.044	0.297

CTEffect = Perceived Cognitive Training Effectiveness, CogDef = Perceived Cognitive Deficit, SE = Standard Error.

Are the investments individuals reported being willing to make sensible in terms of time invested versus time gained in functional independence later in life? We present a brief exploration of this question. To gain one week of functional independence participants were willing to devote an average of 10.91 minutes a day to cognitive training programs. To examine cost relative to benefit, a training time period must be specified. If participants were to, for example, devote 10.91 minutes a day using training for 5 years, they would devote 19,911 minutes, or 331 hours to training. Because a week contains only 168 hours, this means that over a five-year period, time spent training would be much greater than the amount of functional independence gained from training later in life. On the other hand, participants reported being willing to devote 41.99 minutes a day to gain 3 years of functional independence. If participants did devote 41.99 minutes per day to cognitive training for 5 years, they would devote a total of 76,632 minutes or 1,277 hours to training. As three years of independence is equivalent to 26,280 hours, the gain in this case far exceeds the time commitment.

Figure 2 depicts the amount of time (in hours) participants were willing to invest in cognitive training as a function of hours of hypothetical independence gained, assuming participants engaged in daily training for certain periods of time (5, 10, 20 years). As mentioned previously, if an individual engaged in 10.91 minutes of cognitive training every day for five years, that would amount to 331 total hours of investment in training over that 5-year period. Over 10 years, 10.91 minutes of daily training would represent 662 hours invested. Projected out over 20 years, 10.91 minutes of daily training would represent 1,324 hours of invested time. Hours of investment were calculated and plotted in this way based on the average amount of time participants reported being willing to engage in training each day, as a function of hours of benefits obtained. Points falling below the black dashed line in Figure 2 represent more time invested than gained. Except for the very smallest potential gains (1 week to 2 months), the reported investments participants were willing to make did not exceed the hours of independence gained, even assuming a training period of 20 years.

Figure 2.

For Study 1, the amount of time participants reported being willing to invest in cognitive training compared to hypothetical time gained in independence later in life, calculated under the assumption that participants engaged in daily training for 5, 10, and 20 years. All times were converted to hours. Points below the dotted line represent more time invested than time gained through prolonged functional independence.

Discussion

Overall, participants were willing to devote a significant amount of time each day for varying levels of prolonged functional independence. For greater benefits they were willing to devote more time training. Willingness to devote time to cognitive training was influenced by individual difference factors, most significantly by individuals’ perceived effectiveness of cognitive training. Individuals’ perceptions of their own cognition as well as openness to experiences also significantly influenced their devotion to cognitive training. It is interesting to note that perceptions of cognitive training still strongly related to willingness despite participants being told to assume the product being asked about worked. For the very smallest gains in independence, there was some evidence that individuals may have been willing to invest more time in cognitive training compared to hypothetical time gained in functional independence later in life.

STUDY 2 – LAB SAMPLE

This study analyzed data collected from a group of 120 older adults (64+ years old) who completed many of the same or similar measures as participants in Study 1. The primary aim of this study was to understand factors related to adherence to technology-based cognitive interventions. Participants completed a variety of surveys and cognitive measures and were then assigned to engage with a tablet-based cognitive training software program within their own home for 4.5 months while adherence was monitored. This program (Mind Frontiers; Aptima, Inc.) consisted of gamified cognitive exercises, including N-back working memory, spatial reasoning, and task-switching activities (see Souders, Boot, Blocker, Vitale, Roque, & Charness, 2017, for more information about the software program). Data presented here were derived from the baseline sessions of this ongoing study before the intervention began. It should be noted that these are exploratory analyses as the main focus of the study was predicting adherence to training over time.

Method

Participants

One-hundred and twenty older adults were recruited from the Tallahassee community (M_age = 72 years, SD = 5.96, 80 women). For the entire 4.5-month study, participants were compensated a total of $200 dollars. In order to participate, participants had to be 64 years of age or older, report no major deficits with their vision, hearing or physical mobility, and had to be able to recall 6 or more details from a story from the Wechsler Memory Scale (Wechsler, 1997). The sample was 84% white, 13% black, and 3% American Indian/Alaska Native, Multi-Racial, or “other.”

Measures

At baseline, before participants engaged in cognitive training, the Brain Training Independence survey was administered. Instead of being administered online with a slider to indicate daily number of minutes participants were willing to invest in training, participants simply reported numerically the number of hours and minutes for each question using a paper version of the survey (unlike Study 1, participants could report more than 100 minutes if they chose to do so). Also administered were the same general self-efficacy measure, an abbreviated five-question version of the perceived deficit scale, the same demographic questions, and other measures that did not overlap with Study 1. Instead of using a single item, belief in cognitive training efficacy was measured using a survey developed by Rabipour and Davidson (2015). Across a variety of cognitive domains, participants were asked to rate how successful they thought cognitive training would be at improving performance (on a scale of 1–7, with 1 corresponding to “completely unsuccessful” and 7 corresponding to “completely successful”). A score was derived by averaging responses across all domains.

Results

First, we examined the amount of time participants were willing to engage in cognitive training as a function of potential gains in independence. As can be seen in Figure 3, there was a strong positive relationship between hypothetical independence gains and the amount of time participants were willing to invest. Participants reported much greater willingness to devote time to daily training compared to Study 1, possibly due to the open-ended nature of the response format of this study. On average, participants were willing to devote between about 43 minutes (for 1 week of independence) and 160 minutes (for 3 years of independence) each day to training, though individual participants were willing to invest much more time (up to 525 minutes each day). It is unlikely that differences in reported time are due to the large difference in age range between studies. A small study that collected data online from 21 older adults in Tallahassee (M_age = 69 years, SD = 5.96 years) using the same online survey and slider response as Study 1 found that daily minutes devoted to cognitive training ranged from an average of 18.6 minutes for 1 week of independence to 39.9 minutes for 3 years of independence. These are comparable responses to Study 1, suggesting the response mode of Study 2 was likely related to the large differences in reported times observed.

Figure 3:

The mean number of minutes participants were willing to devote to cognitive training as a function of gains in independence for Study 2. Error bars = +/− 1 SEM.

Next, we explored the best predictors of willingness to engage in cognitive training. We conducted a regression analysis with total amount of training time (minutes summed across 7 questions) as the criterion variable, and then age, sex (−1 female, 1 male), self-efficacy, perceived cognitive deficit, and belief in the efficacy of cognitive training as predictors. This model accounted for 9.7 percent of the variance in cognitive training time (F(5, 114) = 2.44, p < .05). Similar to the previous study, the strongest predictor of training time investment was the perceived effectiveness of cognitive training. No other predictors were significant, though there was a trend for willingness to be associated with lower self-efficacy (p =.081). Perceived cognitive deficit was not a significant predictor in the regression, though the bivariate correlation with cognitive training time approached significance (r(118) = .17, p = .06). For the regression analysis, again, the Q-Q plot indicated violation of the assumption of normally distributed residuals. Square root transformation of non-dichotomous data reduced this problem, and results did not meaningfully change (see OSF page for this supplementary analysis, https://osf.io/25czj/?view_only=100c1c46f2e04142a0e1a2efc5da41ba).

It is notable that, compared to Study 1, the amount of variance explained in Study 2 was substantially lower (21.6% for Study 1 vs. 9.7% for Study 2). To better understand why this might be, we conducted an analysis similar to that employed by Study 2, but using Study 1 data (Table 3). This model still accounted for 18.5% of the variance in willingness to engage in cognitive training (F(5, 288) = 13.04, p < .001). Thus, the drop in variance explained does not appear to be strongly related to a more limited number of predictors. It may instead relate to differences in the nature of the samples (e.g., online vs. community-based), or perhaps, the measures (Study 1 and 2 used different measures of cognitive training efficacy, and different methods for participants to enter the number of minutes they were willing to engage in daily cognitive training). It is interesting that when predictors such as health and personality are dropped from the analysis of Study 1 data, age becomes a significant predictor of cognitive training willingness (consistent with our original prediction).

Table 3:

Regression predicting amount of time participants were willing to engage in daily cognitive training for Study 1, using a similar analysis approach to Study 2.

Coefficients
Model	Unstandardized	SE	Standardized	t	p
1 (Intercept)	−22.380	65.008		−0.344	0.731
Age	1.547	0.719	0.119	2.151	0.032
Sex	−4.918	8.653	−0.032	−0.568	0.570
CogDef	1.697	0.747	0.131	2.271	0.024
CTEffect	71.685	10.258	0.380	6.989	< .001
Self-Efficacy	−0.401	1.646	−0.014	−0.244	0.808

CTEffect = Perceived Cognitive Training Effectiveness, CogDef = Perceived Cognitive Deficit, SE = Standard Error.

Returning to Study 2 data, Figure 4 depicts the amount of time participants, on average, were willing to invest in training as a function of hours of independence gained, assuming participants engaged in training for either 5, 10, or 20 years. As a reminder, points falling below the black dashed line represent more time invested than gained. If participants were to devote the amount of time they reported being willing to engage in cognitive training for 5 years, they would invest more time than they would gain for benefits of 1 week, 1 month, and 2 months. For a longer training period of 10 years, the same is true for a gain of 6 months of independence. For larger gains in independence, the amount of gain far exceeded the amount older adults were willing to invest.

Figure 4.

For Study 2, the amount of time participants reported being willing to invest in cognitive training compared to hypothetical time gained in independence later in life, calculated under the assumption that participants engaged in daily training for 5, 10, and 20 years. All times were converted to hours. Points below the dotted line represent more time invested than time gained through prolonged functional independence.

Discussion

Similar to Study 1, participants reported being willing to devote a significant amount of time each day to gain varying levels of prolonged functional independence. The response method of Study 1 (which allowed a maximum response of 100 minutes) may have resulted in an underestimate of participants’ true willingness to engage in cognitive training. The perceived efficacy of the effectiveness of cognitive training was again the best predictor of willingness to engage in daily training. Trends were also observed suggesting that self-efficacy and perceived cognitive deficit may also play a role. As in Study 1, for greater benefits participants were willing to devote more time training, but a similar pattern was found that for relatively small gains participants may have been willing to invest more time in daily training than benefit received.

General Discussion

Across two studies we found that people reported being generally willing to devote a significant amount of time each day to cognitive training programs assuming these programs are effective, and that they were sensitive to differing levels of gained functional independence. Substantial differences in terms of the amount of time participants were willing to devote to cognitive training were observed between the two studies. As Study 2 did not place a limit on the amount of time participants could report, these reports are likely the better estimates of participants’ willingness to engage in daily cognitive training for different amounts of gain.

Both studies found that perceived cognitive training efficacy was the best predictor of willingness to engage in daily training. This is somewhat surprising since participants were explicitly told to assume cognitive training was effective in the survey’s instructions (e.g., “Imagine that there is a proven brain training product on the market. Imagine that the evidence conclusively demonstrates that playing these games every day can allow people to live independently longer as they age.”). Individual questions also started with the phrase “If effective brain games were discovered….,” further promoting an assumption of effectiveness. Despite these cues, the effect of perceived cognitive training efficacy persisted, suggesting a powerful role in shaping individuals’ willingness to adopt brain training.

The association between perceived cognitive training efficacy and willingness to engage in daily training is consistent with the Health Belief Model, in that it proposes that health-related behaviors are shaped by a belief that a threat would be reduced by following a health recommendation (Rosenstock, Strecher & Becker, 1988), and the Protection-Motivation Theory, that posits that health behaviors are shaped by the belief that a recommended response is effective (Rogers, 1983). This concept is captured by the “perceived usefulness” factor of the Senior Technology Acceptance Model (Chen & Chan, 2014). As a form of gerontechnology, older adults who perceive cognitive training as useful (effective) will be more willing to adopt it.

Less powerful and less consistent predictors were perceived cognitive deficit (Study 1 only) and a trend for self-efficacy to be predictive (Study 2). Interestingly, this result was in the opposite direction than the result reported by Mehegan, Rainville, and Skufca (2017), who found a greater percentage of people reporting excellent/very good cognitive functioning among individuals who reported being most willing to engage in cognitive training. However, the effect of perceived cognitive deficit is generally consistent with models that propose that an individual needs to perceive a threat to exist, and that they are vulnerable to that threat, in order for health behaviors to occur. In terms of STAM, perceptions of cognitive deficit may enhance the perceived usefulness of training. Contrary to predictions, age alone was not a significant predictor of willingness to engage in cognitive training. Openness to experience was predictive of willingness to engage in brain training in Study 1, but was not assessed in Study 2. Finally, contrary to predictions, previous experience with friends or family members with dementia, and with cognitive training products, were not important factors shaping willingness to adopt. In Study 1, 37% of participants reported experience with dementia, and 20% reported experience with a cognitive training program, suggesting sufficient variability to detect an effect.

An interesting question is whether game-based cognitive training is “worth it,” that is, whether the gains obtained from training subjectively outweigh the time and effort to obtain those gains. We presented exploratory, and admittedly somewhat speculative, analyses to provide initial data on the tradeoffs individuals are willing to make in terms of time invested. As the potential benefits of cognitive training, and the nature of the training needed to achieve those benefits, become clearer in the future, individuals will be in a better position to make decisions about how to allocate their time. Here we focused on benefits in terms of hypothetical extensions of independence. However, it is possible that benefits beyond extended functional independence may shape older adults’ decisions about whether or not to adopt and adhere to cognitive training. The loss of independence, for example, carries an emotional and financial cost for the individual and their family. These and other factors may be important considerations for older adults when they are weighting the pros and cons of engaging in cognitive training.

There are limitations of the reported studies that need to be considered. In the framework of the Theory of Reasoned Action (Fishbein & Ajzen, 1975), our primary measure of interest is more similar to the construct of “behavioral intention” than it is to behavior. However, data likely provide insight into the factors that shape the initial decision to adopt cognitive training. There are also limitations related to the samples tested. In Study 1, older adult MTurk users are likely not representative of older adults in general, and in Study 2, tested individuals were already part of a cognitive training study. Bias may have had an important impact on reported willingness to engage in cognitive training (though see the Study 1 results section for evidence that Study 1 participants were similar to a small group of community participants in their willingness to engage in cognitive training). In terms of sampling, our sample diversity did not permit comparisons among different racial and ethnic groups, which may be important with respect to willingness to engage in specific cognitively stimulating activities (Friedman et al. 2011; Mehegan, Rainville, & Skufca, 2017). Finally, while multiple attention checks were incorporated into the MTurk survey to ensure high quality data, we are aware that even after disqualifying responses from duplicate IP addresses, there is still the possibility of data contamination as savvy MTurk workers are now using virtual private servers (VPS) to conceal their physical location, allowing them to submit multiple responses from the same device without detection (Dennis, Goodson & Pearson, 2018).

As noted in the introduction, there are two important questions that need to be addressed regarding cognitive training, 1) is the training effective, and 2) will those who might benefit adopt and adhere to training? Unless the answer to the second question is “yes,” potentially effective interventions might fail. This suggests the necessity of a dual approach of understanding both how to design interventions that can change cognition meaningfully, and how to design interventions that maximize adoption and promote long-term adherence.

Table 2:

Regression predicting amount of time participants were willing to engage in daily cognitive training for Study 2.

Coefficients
Model	Unstandardized	SE	Standardized	t	p
1 (Intercept)	934.546	1103.852		0.847	0.399
Age	−8.249	10.547	−0.074	−0.782	0.436
Sex	4.484	61.525	0.007	0.073	0.942
CogDef	23.219	20.623	0.105	1.126	0.263
CTEffect	200.201	80.798	0.229	2.478	0.015
Self-Efficacy	−28.674	16.283	−0.164	−1.761	0.081

CTEffect = Perceived Cognitive Training Effectiveness, CogDef = Perceived Cognitive Deficit, SE = Standard Error.

Appendix A. Brain Training and Independence Survey

As we age, our ability to remember information, react quickly, and solve problems can worsen. A number of brain training products are marketed to prevent or slow these declines. These products usually take the form of simple games that can be played on a computer or mobile device. The ultimate goal of these products is to improve brain functioning so that as we age we can continue to perform the important everyday tasks required to live independently (for example, driving and managing finances).

Imagine that there is a proven brain training product on the market. Imagine that the evidence conclusively demonstrates that playing these brain games every day can allow people to live independently longer as they age. In one study, for example, people who engaged in brain training were able to live independently in their own homes for 6 months longer compared to a control group.

This survey will ask you about the amount of time you would be willing to engage in brain training each day in order to obtain benefits. For example:

• Would you be willing to engage in brain training 6 hours every day, starting now, to gain an additional 2 weeks of independent living later in life?

• Would you be willing to engage in brain training 5 minutes every day, starting now, to gain an additional 6 months of independent living later in life?

These are the types of tradeoffs we would like you to consider. Please answer the following questions assuming that the brain training product works. Think carefully about how much time you would be willing to invest each day, starting now, to receive benefits later in life.

a) If effective brain games were discovered that allowed people to live independently for 1 week longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

b) If effective brain games were discovered that allowed people to live independently for 1 month longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

c) If effective brain games were discovered that allowed people to live independently for 2 months longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

d) If effective brain games were discovered that allowed people to live independently for 6 months longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

e) If effective brain games were discovered that allowed people to live independently for 1 year longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

f) If effective brain games were discovered that allowed people to live independently for 2 years longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)

g) If effective brain games were discovered that allowed people to live independently for 3 years longer, I would be willing to play these games: ______ minutes every day, starting today.

(Slider from 0 to 100 minutes)