Skip to main content
NCBI home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2020 Dec 1;76(9):1777–1787. doi: 10.1093/geronb/gbaa213

Social Support Buffers Against Cognitive Decline in Single Mild Traumatic Brain Injury With Loss of Consciousness: Results From the Canadian Longitudinal Study on Aging

Marc Bedard 1,2, Vanessa Taler 1,2,✉
Editor: Laura Zahodne
PMCID: PMC8557842  PMID: 33254227

Abstract

Objectives

We investigated rates of cognitive decline at 3-year follow-up from initial examination in people reporting mild traumatic brain injury (mTBI) with loss of consciousness (LOC) more than a year prior to initial examination. We examined the role of social support as predictor of preserved cognitive function in this sample.

Method

Analyses were conducted on 440 participants who had self-reported LOC of <1 min, 350 with LOC of 1–20 min, and 10,712 healthy controls, taken from the Canadian Longitudinal Study on Aging (CLSA), a nationwide study on health and aging.

Results

People who reported at baseline that they had experienced mTBI with LOC of 1–20 min more than a year prior were 60% more likely to have experienced global cognitive decline than controls at three-year follow-up. Cognitive decline was most apparent on measures of executive functioning. Logistic regression identified increased social support as predictors of relatively preserved cognitive function.

Discussion

mTBI with longer time spent unconscious (i.e., LOC 1–20 min) is associated with greater cognitive decline years after the head injury. Perceived social support, particularly emotional support, may help buffer against this cognitive decline.

Keywords: CLSA, Cognitive reserve, Mild cognitive impairment, Social interaction, Social support

Although cognitive dysfunction following mild traumatic brain injury (mTBI) has long been believed to remit within the acute period (i.e., the first 3 months post-injury), a growing body of literature has indicated that some individuals continue to experience cognitive impairment long after the head injury (Karr et al., 2014; McInnes et al., 2017). Using a large sample from the Canadian Longitudinal Study on Aging (CLSA), we recently reported that people who had experienced mTBI with loss of consciousness (LOC) more than a year prior were more likely to be cognitively impaired than healthy controls on measures of executive functioning and verbal memory (Bedard et al., 2018, 2020).

Recent research has suggested that mTBI may lead to accelerated brain aging in some cases (Bigler & Stern, 2015; Cole et al., 2015; Santhanam et al., 2019; Tremblay et al., 2019), possibly due to secondary neuropathological changes that are initiated by the primary injury (Bigler & Stern, 2015; Gavett et al., 2011). Interestingly, links have been made between mTBI and associated diffuse white matter cortical thinning in frontal brain regions (Ross et al., 2012; Santhanam et al., 2019; Zhou et al., 2013) and tracts (Tremblay et al., 2019), similar to that found in aged brains (Cole et al., 2015; Santhanam et al., 2019, Tremblay et al., 2019). These findings open the possibility that some of the brain changes following mTBI may be progressive, increasing the risk of early-onset neurodegeneration (Cole et al., 2015). Current evidence indeed indicates that early-life mTBI contributes to greater neurological decline in older adulthood compared to late-life mTBI (Tremblay et al., 2019).

However, evidence has surfaced to support the notion that cognitive outcomes from mTBI are likely to differ as a function of cognitive reserve (Mathias & Wheaton, 2015). Cognitive reserve is typically indexed by measures of intelligence, education or occupational attainment (Bigler & Stern, 2015), and cognitive reserve theories hold that the brain can withstand brain damage either through the use of pre-existing cognitive processes, or through enlisting compensatory strategies (Stern, 2002, 2009). Thus, varying levels of cognitive reserve may explain findings indicating that some people exhibit greater cognitive decline than others after experiencing brain damage. Cognitive reserve has been associated with reduced executive dysfunction 3 months post-mTBI when examined longitudinally (Oldenburg et al., 2016; Stenberg et al., 2020). Similar findings were also reported over a 12-month follow-up period post-TBI in a severity-agnostic investigation involving mTBI participants (Steward et al., 2018); although recovery was most apparent for people with severe TBI, remittance of cognitive deficits as a function of greater cognitive reserve was also observed in mTBI. Similarly, in a cross-sectional study, cognitive reserve was found to mitigate deficits in abstraction, reasoning, and problem solving following TBI, within the first year following head injury (Donders & Stout, 2019).

In addition to the most common measures of cognitive reserve such as premorbid intelligence, education and occupational attainment, a role has been proposed for other later life exposures, including social engagement (Scarmeas et al., 2001; Stern, 2009). That is, increased social contact in later life may attenuate the effects of neurological changes, providing neurocognitive buffering in the face of neurodegeneration. Activities that involve not only social contact but also exposure to greater perceived levels of social support (Evans et al., 2018; Fleck et al., 2019) have also been suggested to confer cognitive reserve against neurodegeneration. Although conventionally different from originally postulated indices of cognitive reserve, these lifetime exposures are not conceptually dissimilar on a neurocognitive level: environmental enrichment has been shown to prevent or slow the accumulation of brain pathology in animal models (Lazarov et al., 2005) and imaging studies have shown the same findings in humans (Anatürk et al., 2018). However, the neuroprotective effects of social contact may be limited to a certain stage of neurodegeneration, because among those who were later diagnosed with Alzheimer’s disease, greater engagement in diverse leisure activities even prior to dementia onset were associated with much more rapid neurocognitive decline (Helzner et al., 2007). Although social contact has been shown to be a possible protective factor long after more severe cases of TBI (Levi et al., 2013; Rassovsky et al., 2015), to our knowledge these links have not been examined among people with mTBI.

The present study extended our previous investigations (Bedard et al., 2018, 2020) by re-examining cognitive outcomes three years after baseline assessment in people who had reported experiencing a single mTBI with varying levels of LOC more than a year prior to initial testing. The association between mTBI and cognitive decline over time remains unclear, particularly when LOC was experienced. The present investigation examined these links using data from the CLSA (Raina et al., 2009, 2019), a large population-stratified study of aging. We identified cognitive decline through the use of reliable change indices (RCIs), a set of calculations to determine true change across testing sessions (i.e., controlling testing variability encountered over multiple assessments). Moreover, we investigated the possible protective influence of social support in participants with mTBI who exhibited cognitive decline.

Materials and Procedures

The data used for the present study were from the CLSA, a large ongoing 20-year study examining health transitions and trajectories in a national stratified random sample of over 50,000 male and female Canadian residents aged 45 to 85 at baseline (Raina et al., 2009, 2019). Ethical review of the CLSA protocol was conducted by the Ethical, Legal, and Social Issues Committee, falling under the jurisdiction of the Canadian Institutes of Health Research (CIHR), and additional research ethics board approval was received for each research site prior to data collection. We used data from the baseline (T1) and the first 3-year follow-up (T2) waves, including demographic and clinical information. The initial sample included two groups, one of 30,097 people (Comprehensive Cohort) who were evaluated through 90-min in-home interviews, as well as in-depth physical and cognitive assessments conducted at one of 11 data collection sites across Canada. The remaining 21,241 participants (Tracking Cohort) were evaluated through phone interviews. The present study uses baseline (T1) and follow-up (T2) data from participants in the Comprehensive Cohort only, because participants in the Tracking Cohort were not asked to report prior TBI. Our institutional research ethics boards provided approval for the present study.

Participants

Detailed information on the sampling frame is provided in Raina et al. (2009). Briefly, CLSA participants include adults between the ages of 45 and 85 years who are fluent in English and/or French. Specific to the secondary analyses of the present investigation, participants were excluded from the larger CLSA dataset if they were ever diagnosed with a neurological disorder, had a cerebrovascular accident, or reported experiencing a concussion or other brain injury in the past 12 months, as were people who reported having experienced multiple injuries, leaving only participants who reported having lost consciousness as a result of a single lifetime head injury that occurred more than 12 months prior to T1 assessment. The same inclusionary criteria were applied at both T1 and T2. For this investigation, mTBI participants who experienced LOC were divided into two groups based on length of LOC: less than one minute (LOC < 1 min) or one to 20 min (LOC 1–20 min). In keeping with mTBI classification guidelines (which state that LOC must be less than 30 min; American Congress of Rehabilitation Medicine, 1993), and as we have done previously (Bedard et al., 2018, 2020) participants who reported losing consciousness for more than 20 min were excluded from our analyses.

It should be noted that the CLSA does not examine posttraumatic amnesia or information for the calculation of a Glasgow Coma Scale score, and the specific date of brain injury is not known: the date of brain injury is recorded as having occurred more than 12 months ago. Finally, in addition to the mTBI groups, we included people who had never experienced a brain injury as our comparison group. Only participants for whom data were available for all the variables of interest were included.

Measures

All participants provided informed consent prior to completing the questionnaires and neuropsychological assessments described below. Demographic and clinical information, including age, education level, sex, marital status, and relative date of brain injury (i.e., more than 12 months prior) were self-reported by participants using questionnaires administered through structured in-person interviews. Included in these analyses are a number of variables selected as covariates, notably: (1) age, because older adults (relative to younger adults) tend to be more susceptible to declines in cognitive functioning (Murman, 2015); (2) gender, given that men may exhibit steeper cognitive decline relative to women (McCarrey et al., 2016); (3) educational attainment, because higher educational levels have been associated with greater cognitive abilities (Wilson et al., 2009); (4) language of administration, which may also affect results, perhaps due to lack of psychometric equivalence on language-based tasks (Pekkala et al., 2009); and (5) depressive levels, because depression has been related to lower cognitive scores (Rock et al., 2014).

Traumatic brain injury

Identification and categorization of the mTBI sample was carried out with the Brief Traumatic Brain Injury Screen (BTBIS; Schwab et al., 2007), a short self-report TBI screening tool. The BTBIS records LOC as having occurred for less than 1 min (LOC < 1 min), between 1 and 20 min (LOC 1–20 min), or more than 20 min, in addition to the number of lifetime TBIs.

Depression

The Center for Epidemiologic Studies Short Depression Scale (CES-D10; Andresen et al., 1994), a 10-item 4-point Likert rating scale, was used to assess depressive symptoms. Items are summed, with a total range from 0 to 30, higher scores indicating greater depressive symptomatology. Clinically, a score of 10 or more suggests significant depressive symptomatology.

Social support

The MOS Social Support Survey (Sherbourne & Stewart, 1991) was used to assess levels of perceived social support. The measure contains 19 items rated on a 5-point Likert scale, from 1 (none of the time) to 5 (all of the time), resulting in a total score between 19 and 95, with higher scores indicative of greater perceived levels of social support. Items are divided into five subscales: (1) emotional support (e.g., positive affect expressions and empathic understanding); (2) informational support (e.g., advisory, guidance, or commentary); (3) tangible support (e.g., providing instrumental aid, either material or behavioral); (4) positive social interaction (e.g., having others available to engage with); and (5) affectionate support (e.g., expressing love and admiration).

Neuropsychological assessment

Participants completed a standardized neuropsychological battery consisting of the Animal Fluency Test (AFT; Rosen, 1980), Controlled Oral Word Association Test (COWAT; Lezak et al., 2004), Mental Alternation Test (MAT; Teng, 1994), and the Victoria Stroop Test (Spreen & Strauss, 1998). The Stroop interference score was derived by dividing task time of the Color-Word trial by the completion time of the Dot trial (Stroop; Strauss et al., 2006). An abbreviated Rey Auditory Verbal Learning Test (RAVLT; Rey, 1964) was administered, consisting of one immediate trial and one delayed recall trial, administered after 5 min. Prospective memory was assessed with the Miami Prospective Memory Test (MPMT; Hernandez Cardenache et al., 2014), which allows for the calculation of a prospective memory score (range: 0–18), as well as a time-based (range: 0–9) and event-based (range: 0–9) subscales, with higher scores indicative of greater prospective memory functioning (for further detail, see Bedard et al., 2018).

Statistical Analyses

Statistical analyses were performed using SPSS version 26 (Armonk, NY: IBM Corp.). All continuous variables were normally distributed as checked with Q-Q plots and determining whether skew statistics remained between −2 and 2 (Curran et al., 1996), with the exception of the Stroop interference and MPMT scores, which were positively skewed and subsequently transformed using logarithmic and inverse functions, respectively. Although skew statistics were acceptable, the presence of negative skew through Q-Q plots was observed on the social support subscales, and subsequently logarithmically transformed. Given heterogeneous sampling probabilities across recruitment locales, which resulted in skewed point-estimates, the CLSA has provided inflation weights (see CLSA, 2017) for descriptive purposes (referred to as trimmed weights), recalibrated by national census data, and weights for use in regressions (referred to as analytic weights) rescaled based on provincial sampling. These weights were used to generate results. A significance level of p <.05 was used for all analyses.

Group differences on continuous demographic and clinical data were analyzed with univariate analysis of variance (ANOVA), contrasts corrected with Bonferroni procedure. Categorical data were analyzed with Kruskal–Wallis tests, and pairwise comparisons were conducted using Mann–Whitney U tests with Bonferroni corrections. Repeated measures analyses of covariance (ANCOVA) were run, with time-point (T1 and T2) and neuropsychological test scores as within-subjects factors, and group (LOC < 1 min, 1–20 min, and controls) as the between-subject factor, while holding age, education, sex, and testing language and depression scores at T2 as covariates. Follow-up univariate tests were then conducted with Bonferroni-corrected pairwise comparisons.

Raw neuropsychological test scores were adjusted by age and education. RCIs were then calculated by subtracting adjusted discrepancy test scores (i.e., T2 – T1) by the control group adjusted mean discrepancy test scores (i.e., T2 mean – T1 mean). This allows for further score adjustment based upon control group practice effects. This subtracted term then serves as the numerator, being divided by the standard error the difference of the control group age- and adjusted mean test score differences (for a more detailed explanation with visual mathematical terms, see Duff, 2012). To then identify cognitive decline versus no change, a z-score cutoff of ±1.645 was applied to the calculated reliable change scores, as is standard convention for RCI (Duff, 2012); people scoring ≤−1.645 were identified as having declined, people with reliable change scores of ≥1.645 were identified as having improved, and people with scores between −1.645 and +1.645 were identified as having remained stable. Bonferroni-corrected Kruskal–Wallis tests were used to identify group differences on the dichotomized (declined vs no change/improved) RCIs across individual tests, as well as the reverse (improved vs no change/declined). Similar sets of analyses were run collapsing RCIs across the test scores, but with decline being identified as deteriorating on two or more tests. This approach was taken in order to identify global cognitive decline. Binary logistic regression analyses were conducted to examine the influences of social support on cognitive decline versus no change on the mTBI group with LOC 1–20 min.

Moderation analyses were performed using a macro tool developed by Hayes (2017) for SPSS, wherein a bootstrapping procedure was used to determine 95% confidence intervals (CIs) based on 5,000 resamples. Group was entered as the predictor variable, and the transformed T2 social support subscale scores were included as moderator, and the dependent variable was the calculated sum of the reliable change scores; this approach allows for an overall estimate of change, whether improvement or decline, across all neuropsychological tests. Unstandardized continuous variables were used for the regression analyses (with the exception of reliable change scores which are by calculation standardized). The same covariates as in previous analyses were entered in the modeling, and Bonferroni correction was used to adjust significance testing.

Results

Demographic and Clinical Characteristics

The final sample in the present study included 11,502 participants: 10,712 no head injury controls, 440 mTBI participants with LOC <1 min, and 350 mTBI participants with LOC of 1–20 min. Demographic and clinical data are presented in Table 1. All three groups were similar in terms of age (p = .08) and education (p = .96), depression scores at T1 (p = .35) and T2 (p = .69), and marital status at T1 (p = .65) and T2 (p = .09). A Time × Group interaction was not significant for depression (p = .09) nor marital status (p = .13). Participants differed on sex (p < .01), with the control group having a smaller proportion of males to females compared to people with LOC <1 min (p < .01, OR = 0.60, 95% CI = 0.50–0.71) and people with LOC 1–20 min (p < .01, OR = 0.71, 95% CI = 0.57–0.86). The two mTBI groups had a similar proportion of females to males (p = .21).

Table 1.

Participant Demographic and Clinical Characteristics

Controls (n = 10,712) LOC < 1 min (n = 440) LOC 1–20 min (n = 350)
T1 T2 T1 T2 T1 T2
Age, mean (SD) 61.51 (9.3) 61.2 (8.8) 60.6 (9.4)
Age, range 47–89 47–88 47–88
Sex
 Female (%) 55.3 42.1 44.9
 Male (%) 44.7 57.9 55.1
Education
 < High school (%) 2.6 4.0 2.1
 High school (%) 9.6 5.7 10.4
 College diploma (%) 37.4 40.6 39.4
 University degree (%) 26.5 28.5 22.2
 Graduate degree (%) 23.9 21.3 25.8
Marital status
 Married (%) 78.9 77.0 81.8 80.3 79.4 77.4
 Widowed (%) 4.4 5.6 3.9 4.2 3.3 3.9
 Divorced (%) 7.2 7.0 5.3 5.2 6.7 5.2
 Separated (%) 1.9 2.5 3.3 5.1 2.1 2.1
 Single (%) 7.6 7.9 5.6 5.1 8.5 11.3
Depression, mean (SD) 8.9 (3.1) 8.6 (2.9) 8.8 (2.9) 8.8 (2.8) 9.1 (3.3) 8.6 (2.9)
Social support total 82.8 (11.6) 83.1 (11.7) 83.3 (11.6) 83.4 (11.8) 82.8 (11.8) 82.6 (12.0)
 Tangible 17.3 (3.0) 17.5 (3.0) 17.6 (3.0) 17.7 (2.9) 17.3 (3.0) 17.3 (3.1)
 Affection 13.7 (2.0) 13.7 (2.0) 13.8 (2.1) 13.8 (2.1) 13.6 (2.2) 13.6 (2.2)
 Positive interaction 13.2 (2.0) 13.2 (2.0) 13.3 (2.1) 13.3 (2.0) 13.3 (2.1) 13.2 (2.0)
 Emotional 34.4 (5.5) 34.5 (5.5) 34.5 (5.4) 34.4 (5.6) 34.5 (5.3) 34.4 (5.5)
Open in a new tab

Note: LOC = loss of consciousness.

Total social support levels were similar between groups at T1 (p = .51) and T2 (p = .48), and the Time × Group interaction was not significant (p = .56). When social support subscales were examined separately, no between-group differences were observed at T1 or T2 on any subscale (ps > .1), none of the Time × Group interactions on the social support subscales were significant, nor were there main effects of Time (ps > .1).

The loss of sample from T1 to T2 was influenced by a number of factors, including the fact that analyses were run on complete cases, which resulted in a total reduction of 2,451 controls, 108 participants with LOC <1 min, and 91 participants with LOC 1–20 min, at T1. Of these, 310 controls, 12 participants with LOC <1 min, and 18 participants with LOC 1–20 min were excluded at T2 due to reporting a neurological disorder, cerebrovascular accident, or having reported a new head injury. Moreover, a total of 2,166 controls, 83 participants with LOC <1 min, and 61 participants with LOC 1–20 min were excluded due to incomplete cognitive data, either due to attrition or because the participant failed to complete particular tests. Participants with missing cognitive data were more likely to be older and to report lower levels of education compared to participants with complete cognitive data (ps < .05). Moreover, controls with complete cognitive data were more likely to report lower levels of total social support than the controls with missing cognitive scores (ps < .05).

Longitudinal Neuropsychological Functioning

A Greenhouse−Geisser corrected repeated measures ANCOVA did not reveal a significant three-way Group by Neuropsychological Test Scores by Time-Point interaction, F(5.80, 30243.16) = 0.38, p = .89, and the Group by Neuropsychological Test Scores interaction was not significant, F(4.04, 30243.16) = 0.85, p = .49. However, there was a significant interaction between Time-Point and Neuropsychological Test, F(2.90, 30243.16) = 11.34, p < .01, which Bonferroni-corrected pairwise comparisons indicated was due to elevated performance on the RAVLT immediate and delayed at T2 compared to T1, ps < .05. The main effect of Group was not significant, F(2, 10430) = 0.54, p = .58.

In further examining the influence of missing cases on results, there were a total of 1,864 controls, 67 participants with LOC <1 min, and 55 participants with LOC 1–20 min who had missingness on the covariates (due to attrition and missing data). The repeated measures ANCOVA was re-run by adding three new group levels—that is, the same sample groupings but including those with missingness on the covariates. This repeated measures ANCOVA revealed a significant main effect of group, F(5, 9911) = 13.037, p < .01, which pairwise comparisons indicated as being due to controls with missingness scoring lower on cognitive measures relative to each of the complete case groups, ps < .05. The other pairwise comparisons were not significant (ps > .05).

Reliable Change Indices

Individual-test reliable deteriorations

A distribution of adjusted reliable change scores is presented in Figure 1. When considering the relative dichotomized frequency of deterioration versus no change/improvement between participants, Bonferroni-corrected Kruskal–Wallis analyses found significant group differences on the AFT, χ 2 (2) = 5.89, p = .04, and on the MAT, χ 2 (2) = 6.0, p = .4, with no group differences in the other neuropsychological measures (ps > .05). Compared to controls, the LOC 1–20 min group included a greater proportion of people exhibiting cognitive deterioration from T1 to T2 on animal fluency (p = .04, OR = 1.62, 95% CI = 1.08–2.16), and on the MAT (p = .04, OR = 1.62, 95% CI = 1.01–2.33). People with LOC 1–20 min were similarly more likely to have deteriorated on the MAT (p = .04, OR = 1.11, 95% CI = 1.01–1.21), and on the AFT (p = .04, OR = 1.10, 95% CI = 1.02–1.18), compared to those with LOC <1 min. No differences were observed between control participants and people with LOC <1 min on the AFT or MAT (ps > .05). Groups did not differ on dichotomized improvement rates (improved vs no change/deterioration), or when omnibus group level differences across the three RCI categorizations (improved vs no change vs deterioration) were examined, all ps > .10.

Figure 1.

Figure 1.

Open in a new tab

Distribution of reliable change scores for each test across study groups.

Note: COWAT = Controlled Oral Word Association Test; LOC = loss of consciousness; MAT = Mental Alternation Test; RAVLT = Rey Auditory Verbal Learning Test. Time-Based and Event-Based refer to the respective subscales of the prospective memory test. The reliable change scores were calculated by first subtracting the adjusted discrepancy test scores (i.e., T2 − T1) by the control group adjusted mean discrepancy test scores (i.e., T2 mean – T1 mean), and dividing this by the SE the difference of the control group age- and education-adjusted mean test score differences.

Global reliable deterioration

Omnibus level group differences emerged when further examining reliable decline on at least two cognitive measures, χ 2 (2) = 5.73, p = .04. Mann–Whitney U test pairwise comparisons indicated that people with LOC 1–20 min were more likely to have declined on two or more tests than controls, (p = .03, OR = 1.55, 95% CI = 1.02–2.35). People with LOC <1 min did not differ from people with LOC 1–20 min or from controls (ps > .1). Reliable improvement on at least two cognitive measures was not found to differ between groups, ps > .10 (Table 2).

Table 2.

Neuropsychological Test and Impairment Rates Across Study Groups

Mild traumatic brain injury
Control (n = 10,712) LOC < 1 min (n = 440) LOC 1–20 min (n = 350)
Test, mean (SD) T1 T2 T1 T2 T1 T2
Declarative memory raw scores
 RAVLT immediate 6.23 (1.85) 7.06 (2.12) 6.26 (1.88) 7.14 (2.17) 6.16 (1.77) 7.09 (2.13)
 RAVLT delayed 4.47 (2.14) 5.23 (2.36) 4.42 (2.06) 5.21 (2.42) 4.43 (2.04) 5.17 (2.36)
 RCI deterioration/improvement
  RAVLT immediate (%) 3.7 / 2.6 3.4 / 2.6 4.3 / 3.4
  RAVLT delayed (%) 3.9 / 3.0 3.7 / 2.9 3.0 / 3.6
Executive functioning raw scores
 Stroop interference 2.07 (0.77) 2.05 (0.62) 2.03 (0.49) 2.01 (0.56) 2.12 (0.81) 2.08 (0.51)
 Mental Alternation Test 27.93 (8.48) 27.76 (7.36) 28.11 (8.30) 27.55 (7.65) 28.03 (7.78) 27.40 (7.28)
 COWAT 40.75 (12.42) 41.59 (12.20) 40.55 (11.77) 41.47 (11.66) 40.38 (12.77) 41.14 (12.95)
 Animal Fluency Test 20.72 (5.58) 20.60 (5.20) 21.20 (5.70) 20.78 (5.12) 20.77 (5.13) 21.13 (5.02)
 Event-based PM 8.64 (1.10) 8.74 (0.96) 8.72 (0.98) 8.82 (0.77) 8.65 (1.03) 8.72 (0.98)
 Time-based PM 8.82 (0.64) 8.73 (0.80) 8.78 (0.63) 8.68 (0.93) 8.75 (0.78) 8.68 (0.92)
 RCI deterioration/improvement
  Stroop interference (%) 5.6 / 4.0 5.0 / 3.6 7.3 / 1.6
  Mental Alternation Test (%) 3.1 / 3.9 2.9 / 4.1 5.6 / 4.6
  COWAT (%) 3.9 / 4.5 3.1 / 5.4 4.7 / 5.7
  Animal Fluency Test (%) 3.1 / 5.4 3.2 / 5.9 5.2 / 5.4
  Event-based PM (%) 4.1 / 6.5 2.4 / 7.0 2.8 / 5.3
  Time-based PM (%) 4.0 / 3.0 3.0 / 3.0 4.9 / 3.8
Global RCI decline/improvement
 Two or more tests (%) 4.5 / 8.3 4.5 / 8.7 7.3 / 10.0
Open in a new tab

Note: COWAT = Controlled Oral Word Association Test; LOC = loss of consciousness; PM = prospective memory; RAVLT = Rey Auditory Verbal Learning Test; RCI = reliable change index. Neuropsychological raw scores are presented and RCI rates represent those calculated on age- and education-adjusted test scores. Deterioration, presented before the forward slash, is indexed as an RCI value of ≤−1.645, and improvement, presented after the forward slash, identifies the percentage of RCI values ≥1.645.

Predictors of Cognitive Deterioration

To identify the influences of social support on decline on two or more cognitive tests in people with LOC 1–20 min, a logistic regression was run including the social support subscales as predictors, along with sex, testing language, depression, and marital status at T2. The initial model, χ 2 (8) = 16.99, p < .01, was further refined with the removal of depression, marital status, and the affection subscale due to low predictive discriminability (Wald = 0.04, 0.02, and 0.3, respectively). The final logistic regression was statistically significant against a constant only model, χ 2 (5) = 19.78, p < .001, Nagelkerke R2 = .14. As shown in Table 3, an increase of one point on the emotional support (OR = 0.85, 95% CI = 0.75–0.97) and positive social interaction scales (OR = 0.67, 95% CI = 0.46–0.88) were associated with a 15% and 33% decrease in odds of cognitive deterioration, respectively. Moreover, identifying as male came with a 461% increased odds of exhibiting cognitive deterioration (OR = 4.61, 95% CI = 1.47–12.58). For those who reported LOC <1 min (χ 2 (5) = 11.05, p =.04, Nagelkerke R2 = .05), a single-point increase in emotional support came with a 10% decrease in odds of cognitive decline (OR = 0.90, 95% CI = 0.82−0.98). A single-point increase in emotional support was associated with a 3% decrease in cognitive decline among controls (OR = 0.97, 95% CI = 0.95–0.99), χ 2 (5) = 37.04, p < .001, Nagelkerke R2 = .01. For controls, being male came with a 143% increase in odds of cognitive decline (OR = 1.43, 95% CI = 1.19−1.72).

Table 3.

Logistic Regression Model for Predictors of Cognitive Deterioration on Two or More Tests

Controls LOC < 1 min LOC 1–20 min
OR 95% CI OR 95% CI OR 95% CI
MOS social support
 Tangible 0.98 0.94, 1.02 0.99 0.79, 1.23 0.85 0.70, 1.04
 Positive interactions 1.03 0.96, 1.11 1.09 0.74, 1.61 0.67 0.46, 0.88
 Emotional 0.97 0.95, 0.99 0.90 0.82, 0.98 0.85 0.75, 0.97
Sex
 Female (reference) 1.00 1.00 1.00
 Male 1.39 1.15, 1.66 1.70 0.63, 4.55 4.22 1.47, 12.31
Open in a new tab

Note: LOC = loss of consciousness; MOS = Medical Outcomes Study; OR = odds ratio. p-values <.05 have been bolded.

Moderating Global Cognitive Decline

Because it was of interest to examine whether social support provided a buffer against cognitive decline, a composite sum was generated across the reliable change scores for each test, which are displayed in Figure 1 (the scores prior to being categorized as declined vs improved). Two-way interactions (requested through PROCESS model 1) on the summed reliable change scores were then evaluated using 5,000 resample bootstrapping to determine 95% CIs (Hayes, 2017), with the same covariates entered as in the MANCOVA. We found a significant interaction between group and emotional support, ΔR2 = .01, b = 0.05, t = 1.98, p = .04. As presented in Figure 2, follow-up simple slope analyses indicated that emotional support moderated the relation between group and cognitive change scores, found to occur to a greater extent in participants with LOC 1–20 min (b = 0.07, p = .01) relative to participants with LOC <1 min (b = 0.03, p < .01) or controls (b = 0.02, p < .01). The interactions of group and social support variables on the summed adjusted change scores were not found for the tangible support, informational support, or positive social interactions subscales, ps > .05.

Figure 2.

Figure 2.

Open in a new tab

Emotional support moderating the relation between group and reliable change scores.

Note: LOC = loss of consciousness. The y-axis is the sum of the reliable change scores across the neuropsychological tests.

Discussion

Increasing research interest has focused on understanding long-term neuropsychological function following mTBI. Although the literature has largely supported the prevailing view that cognitive dysfunction, if present, remits by 3 months after head injury, a subset of people with mTBI experience ongoing cognitive impairment (Bedard et al., 2018, 2020; Karr et al., 2014; McInnes et al., 2017). Beyond cross-sectional findings, little work has been devoted to longitudinal analyses in the post-acute period. However, preliminary indications suggest that mTBI may propagate accelerated brain aging (Bigler & Stern, 2015; Cole et al., 2015; Santhanam et al., 2019; Tremblay et al., 2019).

The present study reports on the first follow-up wave from the CLSA; analyses examining cognitive functioning in mTBI at baseline have been published previously (Bedard et al., 2018, 2020). We found that at 3-year follow-up, people who had experienced mTBI and spent 1–20 min unconscious were 62% more likely to exhibit cognitive decline on the MAT (5.6% vs 3.1%), and 62% more likely to decline on animal fluency (5.2% vs 3.1%), when compared to people who had never experienced a TBI. Relative decline was also found on the AFT (11%) and the MAT (10%) when people with LOC of 1–20 min were compared to people who reported LOC of less than one minute (5.2% vs 3.2%, and 5.6% vs 2.9%, respectively). When considering global cognitive decline, indexed as having declined on two or more neuropsychological measures, people with LOC 1–20 min were 55% more likely to experience cognitive decline (7.3% vs 4.5%) compared to control participants. Notably, none of the groups differed with respect to rates of reliable improvement.

These data therefore provide neuropsychological evidence in line with the notion that mTBI is associated with accelerated neurocognitive decline (Bigler & Stern, 2015; Cole et al., 2015; Santhanam et al., 2019; Tremblay et al., 2019), as suggested by imaging studies, and provide evidence for the importance of LOC as a clinical consideration. Consistent with prior findings of diffuse axonal thinning and anisotropy within the frontal lobe (Ross et al., 2012; Santhanam et al., 2019; Tremblay et al., 2019; Zhou et al., 2013), the cognitive decline observed among people who have experienced mTBI with LOC between 1 and 20 min was apparent on measures of executive functioning. The decline occurred on tasks assumed to require executive control, set-shifting, and inhibition, highly reliant on working memory and processing speed—processes that are subserved by frontal tracts (Sasson et al., 2013). The present data indicate that these associations were stronger in people who experienced unconsciousness for a longer period of time (i.e., 1–20 min).

We also found an effect of social support on cognitive deterioration in mTBI: in people who experienced mTBI with LOC 1–20 min, a single-point increase on emotional support (which ranges from 0 to 40) and positive social interaction (0–15 range) were associated with 15% and 33% decreases in the likelihood of exhibiting cognitive decline, respectively. A role for social support being associated with greater cognitive functioning was further evidenced in the interaction between emotional support and group on summed reliable change scores: although all groups were more likely to perform better at T2 (relative to T1) when they had higher levels of emotional support, this relationship was stronger for participants with LOC 1–20 min than for controls or participants with LOC <1 min. These data therefore suggest that social support, particularly emotional support—perceiving that one has others to provide empathic understanding and who can reflect positive affective expressions—is associated with a lower long-term likelihood of experiencing cognitive decline following mTBI. These findings fit with the growing evidence suggesting that social contact may buffer against cognitive dysfunction (Evans et al., 2018; Fleck et al., 2019; Scarmeas et al., 2001; Stern, 2009), and neurological deterioration (Anatürk et al., 2018). Previous research has indicated that social contact can have protective neurocognitive influences in people with severe TBI (Rassovsky et al., 2015; Levi et al., 2013), and the current study extends this literature with findings to suggest that people with mTBI may benefit neuropsychologically from social support.

The present investigation takes advantage of the large sample size of the CLSA to assess long-term cognitive functioning following mTBI with LOC. It should be noted that analyses aiming to identify cognitive decline in the present study were conducted on age- and education-adjusted raw scores, and that the RCIs were calculated by using the standard error of the difference of mean control group performance, accounting for practice effects on repeat testing (Duff, 2012). Using age- and education-adjusted scores while accounting for practice effects is a unique strength of the present study, allowing for robust substantiation that cognitive declines are in fact reliable.

We also note a few limitations to the present study. Although these data suggest that people who experienced mTBI with LOC of 1–20 min are more likely to experience cognitive decline at 3-year follow-up relative to controls, it is also possible that what is observed is due to greater score variability as opposed to definite impairment. However, we contend that if this were the case, the variability should be weighted toward lower scores at follow-up, particularly given that group differences did not emerge when looking at rates of reliable improvement. As has been discussed in our prior papers, the incidence of TBIs across participants is self-reported, increasing the risk of under or over-reporting of mTBI status, and biasing the validity of the length of time spent unconscious. Moreover, the CLSA does not collect data on Glasgow Coma Scale Scores nor information to account for posttraumatic amnesia, making it challenging to assess TBI severity accurately. It is also possible that participants perceived questions related to head injury and concussion differently than intended, and thus some who experienced head injury may have not been identified. Given that the specific date or time since the mTBI is not known, it is not possible to make any definitive assertion with respect to temporality of findings—we know that the initial mTBI occurred more than 12 months before T1 testing, but how long before remains unknown. Premorbid intellectual functioning likewise cannot be assessed.

It is also noteworthy that the dataset for the present investigation made use of complete cases, with listwise deletions to remove cases that had missing data. Additional analyses indicated that participants who were missing cognitive data were more likely to be older, have less formal education, and that the control group was more likely to report lower levels of social support. Similarly, controls who had missing data on the independent variables were also more likely to have lower cognition performance. Notwithstanding the other analyses, caution is advised when interpreting the RCIs. Given that the RCIs were not only adjusted for by age and education, but made use of the control group mean differences and standard errors of the difference between T1 and T2, it is possible that relying on complete cases may have inadvertently led to observing less cognitive decline among controls relative to the other groups.

Conclusions

We found that people who self-reported mTBI with LOC that occurred more than a year prior were more likely to exhibit cognitive decline relative to healthy control participants over a 3-year follow-up period. It is notable that these associations occurred in the context of greater time spent unconscious (i.e., 1–20 min), indicating that length of unconsciousness is an important factor in guiding long-term outcomes. Moreover, in people who exhibited cognitive decline, greater social support, particularly empathic and positive interactions with others, were associated with a lower likelihood of cognitive decline.

Acknowledgments

This research was made possible using the data/biospecimens collected by the Canadian Longitudinal Study on Aging (CLSA). This research has been conducted using the CLSA Baseline Comprehensive Dataset 3.1 and Follow-up 1 Comprehensive Dataset 2.0, under Application Number 190228. The CLSA is led by Drs. Parminder Raina, Christina Wolfson, and Susan Kirkland. Data are available from the CLSA (www.clsa-elcv.ca) for researchers who meet the criteria for access to de-identified CLSA data.

Funding

Funding for the Canadian Longitudinal Study on Aging (CLSA) is provided by the Government of Canada through the Canadian Institutes of Health Research (CIHR; grant number LSA 94473) and the Canada Foundation for Innovation. M. Bedard is a Vanier Scholar, funded by the CIHR.

Conflict of Interest

None declared.

References

  1. American Congress of Rehabilitation Medicine, Mild Traumatic Brain Injury Committee of the Interdisciplinary Special Interest Group. (1993). Definition of mild traumatic brain injury. Journal of Head Trauma and Rehabilitation, 8(3), 86–87. [Google Scholar]
  2. Anatürk, M., Demnitz, N., Ebmeier, K. P., & Sexton, C. E. (2018). A systematic review and meta-analysis of structural magnetic resonance imaging studies investigating cognitive and social activity levels in older adults. Neuroscience and Biobehavioral Reviews, 93, 71–84. doi: 10.1016/j.neubiorev.2018.06.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Andresen, E. M., Malmgren, J. A., Carter, W. B., & Patrick, D. L. (1994). Screening for depression in well older adults: Evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). American Journal of Preventive Medicine, 10(2), 77–84. doi: 10.1016/S0749-3797(18)30622-6 [DOI] [PubMed] [Google Scholar]
  4. Bedard, M., Steffener, J., & Taler, V. (2020). Long-term cognitive impairment following single mild traumatic brain injury with loss of consciousness: Findings from the Canadian Longitudinal Study on Aging. Journal of Clinical and Experimental Neuropsychology, 42(4), 344–351. doi: 10.1080/13803395.2020.1714552 [DOI] [PubMed] [Google Scholar]
  5. Bedard, M., Taler, V., & Steffener, J. (2018). Long-term prospective memory impairment following mild traumatic brain injury with loss of consciousness: Findings from the Canadian Longitudinal Study on Aging. The Clinical Neuropsychologist, 32(5), 1002–1018. doi: 10.1080/13854046.2017.1404644 [DOI] [PubMed] [Google Scholar]
  6. Bigler, E. D., & Stern, Y. (2015). Traumatic brain injury and reserve. Handbook of Clinical Neurology, 128, 691–710. doi: 10.1016/B978-0-444-63521-1.00043-1 [DOI] [PubMed] [Google Scholar]
  7. Canadian Longitudinal Study on Aging (CLSA). (2017). Sampling and computation of response rates and sample weights for the tracking (telephone interview) participants and comprehensive participants.https://www.clsa-elcv.ca/doc/1041
  8. Cole, J. H., Leech, R., & Sharp, D. J.; Alzheimer’s Disease Neuroimaging Initiative . (2015). Prediction of brain age suggests accelerated atrophy after traumatic brain injury. Annals of Neurology, 77(4), 571–581. doi: 10.1002/ana.24367 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Curran, P. J., West, S. G., & Finch, J. F. (1996). The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychological Methods,1(1), 16–29. doi: 10.1037/1082-989X.1.1.16 [DOI] [Google Scholar]
  10. Donders, J., & Stout, J. (2019). The influence of cognitive reserve on recovery from traumatic brain injury. Archives of Clinical Neuropsychology, 34(2), 206–213. doi: 10.1093/arclin/acy035 [DOI] [PubMed] [Google Scholar]
  11. Duff, K. (2012). Evidence-based indicators of neuropsychological change in the individual patient: Relevant concepts and methods. Archives of Clinical Neuropsychology, 27(3), 248–261. doi: 10.1093/arclin/acr120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Evans, I. E. M., Llewellyn, D. J., Matthews, F. E., Woods, R. T., Brayne, C., & Clare, L.; CFAS-Wales Research Team . (2018). Social isolation, cognitive reserve, and cognition in healthy older people. PLOS One, 13(8), e0201008. doi: 10.1371/journal.pone.0201008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fleck, J. I., Arnold, M., Dykstra, B., Casario, K., Douglas, E., & Morris, O. (2019). Distinct functional connectivity patterns are associated with social and cognitive lifestyle factors: Pathways to cognitive reserve. Frontiers in Aging Neuroscience, 11, 310. doi: 10.3389/fnagi.2019.00310 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gavett, B. E., Stern, R. A., & McKee, A. C. (2011). Chronic traumatic encephalopathy: A potential late effect of sport-related concussive and subconcussive head trauma. Clinics in Sports Medicine, 30(1), 179–88, xi. doi: 10.1016/j.csm.2010.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). The Guilford Press. [Google Scholar]
  16. Helzner, E. P., Scarmeas, N., Cosentino, S., Portet, F., & Stern, Y. (2007). Leisure activity and cognitive decline in incident Alzheimer disease. Archives of Neurology, 64(12), 1749–1754. doi: 10.1001/archneur.64.12.1749 [DOI] [PubMed] [Google Scholar]
  17. Hernandez Cardenache, R., Burguera, L., Acevedo, A., Curiel, R., & Loewenstein, D. A. (2014). Evaluating different aspects of prospective memory in amnestic and nonamnestic mild cognitive impairment. ISRN Neurology, 2014, 805929. doi: 10.1155/2014/805929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Karr, J. E., Areshenkoff, C. N., & Garcia-Barrera, M. A. (2014). The neuropsychological outcomes of concussion: A systematic review of meta-analyses on the cognitive sequelae of mild traumatic brain injury. Neuropsychology, 28, 321–336. doi: 10.1037/neu0000037 [DOI] [PubMed] [Google Scholar]
  19. Lazarov, O., Robinson, J., Tang, Y. P., Hairston, I. S., Korade-Mirnics, Z., Lee, V. M., Hersh, L. B., Sapolsky, R. M., Mirnics, K., & Sisodia, S. S. (2005). Environmental enrichment reduces Aβ levels and amyloid deposition in transgenic mice. Cell, 120(5), 701–713. doi: 10.1016/j.cell.2005.01.015 [DOI] [PubMed] [Google Scholar]
  20. Levi, Y., Rassovsky, Y., Agranov, E., Sela-Kaufman, M., & Vakil, E. (2013). Cognitive reserve components as expressed in traumatic brain injury. Journal of the International Neuropsychological Society, 19(6), 664–671. doi: 10.1017/S1355617713000192 [DOI] [PubMed] [Google Scholar]
  21. Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological assessment (4th ed.). Oxford University Press. [Google Scholar]
  22. Mathias, J. L., & Wheaton, P. (2015). Contribution of brain or biological reserve and cognitive or neural reserve to outcome after TBI: A meta-analysis (prior to 2015). Neuroscience and Biobehavioral Reviews, 55, 573–593. doi: 10.1016/j.neubiorev.2015.06.001 [DOI] [PubMed] [Google Scholar]
  23. McCarrey, A. C., An, Y., Kitner-Triolo, M. H., Ferrucci, L., & Resnick, S. M. (2016). Sex differences in cognitive trajectories in clinically normal older adults. Psychology and Aging, 31(2), 166–175. doi: 10.1037/pag0000070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. McInnes, K., Friesen, C. L., MacKenzie, D. E., Westwood, D. A., & Boe, S. G. (2017). Mild traumatic brain injury (mTBI) and chronic cognitive impairment: A scoping review. PLoS One, 12(4), e0174847. doi: 10.1371/journal.pone.0174847 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Murman, D. L. (2015). The impact of age on cognition. Seminars in Hearing, 36(3), 111–121. doi: 10.1055/s-0035-1555115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Oldenburg, C., Lundin, A., Edman, G., Nygren-de Boussard, C., & Bartfai, A. (2016). Cognitive reserve and persistent post-concussion symptoms—a prospective mild traumatic brain injury (mTBI) cohort study. Brain Injury, 30(2), 146–155. doi: 10.3109/02699052.2015.1089598 [DOI] [PubMed] [Google Scholar]
  27. Pekkala, S., Goral, M., Hyun, J., Obler, L. K., Erkinjuntti, T., & Albert, M. L. (2009). Semantic verbal fluency in two contrasting languages. Clinical Linguistics & Phonetics, 23(6), 431–445. doi: 10.1080/02699200902839800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Raina, P., Wolfson, C., Kirkland, S., Griffith, L. E., Balion, C., Cossette, B., Dionne, I., Hofer, S., Hogan, D., van den Heuvel, E. R., Liu-Ambrose, T., Menec, V., Mugford, G., Patterson, C., Payette, H., Richards, B., Shannon, H., Sheets, D., Taler, V., Tuokko, H., Wister, A., Wu, C., & Young, L. (2019). Cohort profile: The Canadian Longitudinal Study on Aging (CLSA). International Journal of Epidemiology, 48(6), 1752–1753j. doi: 10.1093/ije/dyz173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Raina, P., Wolfson, C., Kirkland, S. A., Griffith, L. E., Oremus, M., Patterson, C., Tuokko, H., Penning, M., Balion, C. M., Hogan, D., Wister, A., Payette, H., Shannon, H., & Brazil, K. (2009). The Canadian Longitudinal Study on Aging (CLSA) [Special Issue]. Canadian Journal on Aging, 28(3), 221–229. doi: 10.1017/S0714980809990055 [DOI] [PubMed] [Google Scholar]
  30. Rassovsky, Y., Levi, Y., Agranov, E., Sela-Kaufman, M., Sverdlik, A., & Vakil, E. (2015). Predicting long-term outcome following traumatic brain injury (TBI). Journal of Clinical and Experimental Neuropsychology, 37(4), 354–366. doi: 10.1080/13803395.2015.1015498 [DOI] [PubMed] [Google Scholar]
  31. Rey, A. (1964). L’examen Clinique en psychologie. Presses universitaires de France. [Google Scholar]
  32. Rock, P. L., Roiser, J. P., Riedel, W. J., & Blackwell, A. D. (2014). Cognitive impairment in depression: A systematic review and meta-analysis. Psychological Medicine, 44(10), 2029–2040. doi: 10.1017/S0033291713002535 [DOI] [PubMed] [Google Scholar]
  33. Rosen, W. G. (1980). Verbal fluency in aging and dementia. Journal of Clinical Neuropsychology, 2(2), 135–146. doi: 10.1080/01688638008403788 [DOI] [Google Scholar]
  34. Ross, D. E., Ochs, A. L., Seabaugh, J. M., Demark, M. F., Shrader, C. R., Marwitz, J. H., & Havranek, M. D. (2012). Progressive brain atrophy in patients with chronic neuropsychiatric symptoms after mild traumatic brain injury: A preliminary study. Brain Injury, 26(12), 1500–1509. doi: 10.3109/02699052.2012.694570 [DOI] [PubMed] [Google Scholar]
  35. Santhanam, P., Wilson, S. H., Oakes, T. R., & Weaver, L. K. (2019). Accelerated age-related cortical thinning in mild traumatic brain injury. Brain and Behavior, 9(1), 1–8. doi: 10.1002/brb3.1161 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sasson, E., Doniger, G. M., Pasternak, O., Tarrasch, R., & Assaf, Y. (2013). White matter correlates of cognitive domains in normal aging with diffusion tensor imaging. Frontiers in Neuroscience, 7(23), 1–13. doi:10.3389/fnins.2013.00032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Scarmeas, N., Levy, G., Tang, M. X., Manly, J., & Stern, Y. (2001). Influence of leisure activity on the incidence of Alzheimer’s disease. Neurology, 57(12), 2236–2242. doi: 10.1212/wnl.57.12.2236 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Schwab, K. A., Ivins, B., Cramer, G., Johnson, W., Sluss-Tiller, M., Kiley, K., Lux, W., & Warden, D. (2007). Screening for traumatic brain injury in troops returning from deployment in Afghanistan and Iraq: Initial investigation of the usefulness of a short screening tool for traumatic brain injury. The Journal of Head Trauma Rehabilitation, 22(6), 377–389. doi: 10.1097/01.HTR.0000300233.98242.87 [DOI] [PubMed] [Google Scholar]
  39. Sherbourne, C. D., & Stewart, A. L. (1991). The MOS Social Support Survey. Social Science & Medicine (1982), 32(6), 705–714. doi: 10.1016/0277-9536(91)90150-b [DOI] [PubMed] [Google Scholar]
  40. Spreen, O., & Strauss, E. (1998). A compendium of neuropsychological tests (2nd ed.). Oxford University Press. [Google Scholar]
  41. Stenberg, J., Håberg, A. K., Follestad, T., Olsen, A., Iverson, G. L., Terry, D. P., Karlsen, R. H., Saksvik, S. B., Karaliute, M., Ek, J. A. N., Skandsen, T., & Vik, A. (2020). Cognitive reserve moderates cognitive outcome after mild traumatic brain injury. Archives of Physical Medicine and Rehabilitation, 101(1), 72–80. doi: 10.1016/j.apmr.2019.08.477 [DOI] [PubMed] [Google Scholar]
  42. Stern, Y. (2002). What is cognitive reserve? Theory and research application of the reserve concept. Journal of the International Neuropsychological Society, 8(3), 448–460. [PubMed] [Google Scholar]
  43. Stern, Y. (2009). Cognitive reserve. Neuropsychologia, 47(10), 2015–2028. doi: 10.1016/j.neuropsychologia.2009.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Steward, K. A., Kennedy, R., Novack, T. A., Crowe, M., Marson, D. C., & Triebel, K. L. (2018). The role of cognitive reserve in recovery from traumatic brain injury. The Journal of Head Trauma Rehabilitation, 33(1), E18–E27. doi: 10.1097/HTR.0000000000000325 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A compendium of neuropsychological tests: Administration, norms, and commentary (3rd ed.). Oxford University Press. [Google Scholar]
  46. Teng, E. L. (1994). The Mental Alternation Test (MAT). Department of Neurology, University of Southern California School of Medicine. [Google Scholar]
  47. Tremblay, S., Desjardins, M., Bermudez, P., Iturria-Medina, Y., Evans, A., Jolicoeur, P., & De Beaumont, L. (2019). Mild traumatic brain injury: The effect of age at trauma onset on brain structure integrity. NeuroImage: Clinical, 23, 1–6. doi: 10.1016/j.nicl.2019.101907 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wilson, R. S., Hebert, L. E., Scherr, P. A., Barnes, L. L., Mendes de Leon, C. F., & Evans, D. A. (2009). Educational attainment and cognitive decline in old age. Neurology, 72(5), 460–465. doi: 10.1212/01.wnl.0000341782.71418.6c [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Zhou, Y., Kierans, A., Kenul, D., Ge, Y., Rath, J., Raume, J., Grossman, R. I., & Lui, Y. W. (2013). Mild traumatic brain injury: Longitudinal regional brain volume changes. Radiology, 267(3), 880–890. doi: 10.1148/radiol.13122542 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES