Cognitive restructuring is one of the central therapeutic skills in most Cognitive Behavioural Therapy (CBT). Cognitive restructuring involves identifying, challenging and modifying unhelpful or unrealistic thoughts. It is unclear whether there are age-differences in the ability to learn and benefit from cognitive restructuring as a therapeutic skill across the lifespan. Developmentally-appropriate CBT protocols with children typically include cognitive restructuring skills, similar to adult-focused CBT. Despite evidence suggesting that older adults are able to learn cognitive therapy skills (Johnco et al., 2013a, 2014a, 2015a), there is more variability in the application of CBT with older adults, where cognitive restructuring is often omitted or unnecessarily simplified given concerns about normative age-related changes in cognitive flexibility. There is a need to understand developmental differences in the ability to learn cognitive restricting skills, as well as whether cognitive flexibility plays a similar or differential role in the ability to learn cognitive restructuring skills across the lifespan.
Developmental maturation and changes are important to consider during psychological treatments, particularly in cognitive abilities. Cognitive abilities show significant maturation during childhood and adolescence, typically peaking during adulthood followed by normative declines in some skills during older age (Schaie & Willis, 2010; Tervo-Clemmens et al., 2023). As such, there has been interest in whether these developmental changes in cognitive ability influence the suitability of particular therapeutic skills across the lifespan, with cognitive restructuring being one of the most debated. There has been longstanding consideration about whether children possess sufficient cognitive skills to engage in cognitive restructuring techniques, with particular emphasis on the maturation of abstract thinking skills, as well as self-reflection and insight skills. Despite this, most evidence-based CBT programs for anxiety disorders in children routinely include cognitive restructuring as a core therapeutic skill for children from age 7 (e.g., Cools Kids and Coping Cat; Kendall & Hedtke, 2006; Rapee et al., 2019) even in the context of developmental variability in the constituent cognitive skills required to implement this therapeutic skill. At the other end of the age spectrum, there has been long-standing controversy about whether older adults can learn cognitive therapy skills. This has focused on concerns that normative age-related changes in cognitive skills, such as declines in cognitive flexibility (the ability to switch mental sets and inhibit habitual responding in response to changing needs; Diamond, 2013), may result in older adults being too ‘set in their ways’ or ‘rigid’ to be able to fully utilise cognitive restructuring, which requires the ability to consider alternatives and change beliefs. As a result, there is comparatively more variability in whether cognitive therapy skills are included (Dick et al., 1995; Wuthrich, 2009), simplified or omitted (e.g., use of coping statements, alternative thought generation or thought stopping; Stanley et al., 2004; Stanley et al., 2007; Wetherell et al., 2007) with older adults. However it is unclear whether there are differences in the ability to learn cognitive restructuring skills across the lifespan.
Several studies have shown that cognitive flexibility skills are related to the ability to learn cognitive restructuring in older (Johnco et al., 2013a, 2014a, 2015a; Mohlman, 2013), and younger adults (Holder et al., 2021). These effects have been demonstrated during single-session experimental studies, as well as in treatment contexts that assess pre-treatment cognitive flexibility as a predictor of cognitive restructuring skills at post-treatment. For example, despite those with worse cognitive flexibility showing more initial difficulty learning cognitive restructuring (Johnco et al., 2015a), there is some suggestion that they can become more proficient over the course of treatment with repetition and practice (Johnco et al., 2014a). However this has not been examined in children, and it is unclear whether there are age-differences in the ability to learn and benefit from cognitive restructuring.
In comparison to the literature suggesting cognitive flexibility is important for learning cognitive restructuring, there is more mixed evidence about whether executive function skills (including cognitive flexibility) are related to treatment outcomes. Some studies have suggested that these cognitive skills do not predict treatment outcome in children (Godovich et al., 2020; Moritz et al., 2005), adults (Voderholzer et al., 2013) or older adults with anxiety (Johnco et al., 2014a; Mohlman, 2013), while others have found worse treatment outcome (or paradoxically, better; Hybel et al., 2017) among those with executive dysfunction in children (Flessner et al., 2010; McNamara et al., 2014) and adults with OCD (D’Alcante et al., 2012). A small pilot study found improved outcomes following CBT with adjunctive executive skills training compared to CBT alone in older adults with anxiety (Mohlman, 2008). In addition, there are studies with children (Godovich et al., 2020) and older adults (Mohlman, 2013; Mohlman & Gorman, 2005) that have suggested that treatment outcome depends on the stability of executive dysfunction, with those who show improvement in executive functioning skills over treatment showing superior outcome to those who showed persistent impairment both before and after treatment. However these studies have found that baseline executive dysfunction does not differentiate these differential treatment trajectories of individual patients with executive dysfunction, limiting its utility for treatment personalisation or modification. Overall, there appears to be more mixed findings about the relationship between executive functioning and CBT treatment outcomes, but more consistent evidence suggesting a relationship between executive functioning (and specifically - cognitive flexibility) and cognitive restructuring in both older and younger adults, however this has not been extended to children. It may be that cognitive flexibility skills are relevant for certain therapeutic skills like cognitive restructuring, but not for all CBT skills. As a result, impaired executive function may not necessarily negate a person’s ability to benefit from CBT programs that include multiple therapeutic skills.
Cognitive restructuring is a core skill in CBT, however there is limited understanding about how developmental and age-related cognitive changes influence the learning and application of this specific skill across the lifespan. There is increasing interest in the personalisation of psychological treatments, where individual client characteristics are used to tailor treatment by offering only the most relevant and effective treatment components. There is evidence to suggest that individual differences in cognitive ability may be relevant for treatment personalisation (at least in some age groups), with potential implications for emphasising or de-emphasising cognitive restructuring skills in treatment. However, there is a need to understand its utility and relevance across the age spectrum. This study had three aims. Firstly, this study aimed to examine whether there are age-difference in cognitive restructuring skill acquisition between children, younger adults and older adults. Second, this study aimed to examine whether individual differences in cognitive flexibility plays a role in the ability to learn cognitive restructuring across the lifespan. Third, this study aimed to assess whether cognitive flexibility explains (mediates) any age-differences in cognitive restructuring skills.
Method
Participants
This study included 114 participants with an anxiety disorder, including 35 children (Aged 7–12, M = 9.14, SD = 1.44; 60% Female), 32 younger adults (aged 18–53, M = 23, SD = 7.84; 66% Female) and 47 older adults (Aged 61–78, M = 66.81, SD = 4.43; 53% Female). Participant characteristics are summarised in Table 1. All younger and older adult participants also had a unipolar depressive disorder diagnosis, however this was not required for child participants given the low prevalence of depressive disorders in childhood (Spoelma et al., 2023). Child participants were recruited through community advertising, targeted social media ads, and advertising at a specialist anxiety research clinic. Adult participants were recruited via undergraduate research participation pools and flyers placed around a university. The older adult sample was secondary data analysis from Johnco et al. (2015b), and were recruited via the CBT arm of a randomised controlled trial (Wuthrich et al., 2016). Exclusion criteria included clinically significant suicide risk, concurrent CBT, diagnosis of Oppositional Defiant Disorder, Bipolar Disorder, Substance Use Disorder, Psychosis, Autism Spectrum Disorder, significant learning disorder, intellectual disability, or Dementia.
Table 1
Sample characteristics
Child (N = 35) | Younger Adult (N = 32) | Older Adult (N = 47) | ||
|---|---|---|---|---|
n (%) | n (%) | n (%) | χ2 | |
Sex | 1.25 | |||
Female | 21 (600%) | 21 (65.6) | 25 (53.2%) | |
Male | 14 (40.0%) | 11 (34.4%) | 22 (46.8%) | |
Born in Australia | 32 (91.4%) | 22 (68.8%) | 37 (78.7%) | 5.40 |
Primary Diagnosis | 14.30*** | |||
GAD | 18 (51.4%) | 9 (28.1%) | 17 (36.2%) | |
Social Phobia | 6 (17.1%) | 8 (25.0%) | 3 (6.4%) | |
Specific Phobia | 8 (22.8%) | 3 (9.4%) | 1 (2.1%) | |
Separation Anxiety | 3 (8.6%) | 1 (3.1%) | 0 (0.0%) | |
Major Depressive Disorder | 0 (0.0%) | 5 (15.6%) | 12 (25.5%) | |
PDD | 0 (0.0%) | 3 (9.4%) | 4 (8.5%) | |
Panic Disorder | 0 (0.0%) | 2 (6.3%) | 0 (0.0%) | |
PTSD | 0 (0.0%) | 1 (3.1%) | 1 (2.1%) | |
Agoraphobia | 0 (0.0%) | 0 (0.0%) | 1 (2.1%) | |
OSAD | 0 (0.0%) | 0 (0.0%) | 6 (12.8%) | |
OSDD | 0 (0.0%) | 0 (0.0%) | 2 (4.3%) | |
Comorbid diagnosis | 29 (82.9%) | 32 (100.00%) | 47 (100%) | 14.30*** |
Comorbid anxiety disorder | 29 (82.9%) | 29 (90.6%) | 34 (72.3%) | 4.24 |
Comorbid affective disorder | 0 (0.00%) | 32 (100.00%) | 47 (100%) | 114.00*** |
Measures
Diagnostic Interview
Diagnostic status in younger and older adults was assessed using the Anxiety and Related Disorders Interview Schedule for DSM-5 (ADIS; Brown & Barlow, 2014) and ADIS-IV (Brown et al., 1994) respectively, and children were diagnosed using the Youth Online Diagnostic Assessment (YODA; McLellan et al., 2021). The ADIS is a semi-structured diagnostic interview to diagnose anxiety, affective and related diagnoses. Disorder severity for each diagnosis is rated using a Clinician Severity Rating (CSR) ranging from 0 to 8, with scores ≥ 4 indicative of meeting or exceeding diagnostic thresholds. The ADIS was administered by graduate-level Clinical Psychology students supervised by a registered Clinical Psychologist. Children were diagnosed using the parent-report Youth Online Diagnostic Assessment (YODA; McLellan et al., 2021). The YODA is an online diagnostic assessment used to assess and diagnose anxiety disorders in children and adolescents according to DSM-V (American Psychiatric Association, 2013) criteria. The YODA includes a combination of closed and open questions assessing core diagnostic criteria for anxiety, depressive, oppositional, and attention deficit disorders. The measure was initially computer-scored and then subsequently reviewed jointly by a Provisional Psychologist and registered Clinical Psychologist, alongside clinical intake information to determine the final diagnostic profile and severity ratings for each disorder. Participants receive a dichotomous yes/no rating for the presence of each disorder, and disorder severity for each diagnosis is rated on a clinician severity score (0 = Minimal or absent, 1 = Mild, 2 = Moderate, 3 = Substantial, and 4 = Pervasive). The clinician-reviewed YODA has demonstrated high inter-rater reliability and good construct validity with the ADIS-IV in school-aged and preschool aged children (McLellan et al., 2021; Morgan et al., 2019).
Addenbrooke’s Cognitive Examination—Revised (ACE-R; Mioshi et al., 2006)
The ACE-R was administered to older adult participants only, and is a cognitive screening measure sensitive to early dementia (Mioshi et al., 2006). It assesses five cognitive domains (orientation/attention, memory, verbal fluency, language, and visuospatial skills). There were three participants (6.3%) who scored on the cut-off score (< 82/100) for probable cognitive impairment, but were included in analyses given they did not score below this threshold (Mioshi et al., 2006).
WCST. Wisconsin Card Sorting Test – Computer Version 4 (WCST; Heaton, 2003)
The WCST is a computerized neuropsychological test of cognitive flexibility, including cognitive set-shifting, problem-solving, and the ability to modulate responding based on changing rules and feedback (Spreen & Strauss, 1998). Participants were required to sort cards (differing in colour, shape, and number) shown on a computer screen without explicit instructions. Participants are required to use feedback on their sorting (correct/incorrect) to determine the matching rule (colour, shape, or number). After 10 correct sorts, the matching rule changes without warning and participants are required to update their sorting principles without perseverating back to previous sorting principles. Age- and education-adjusted perseverative error T-scores (WCST-PE) were used in the current study given previous research demonstrating specific relationship with perseveration and cognitive restructuring (Holder et al., 2021; Johnco et al., 2013a, 2014a, 2015a). Higher T-scores reflect better functioning (i.e., less perseveration).
Controlled Oral Word Association Test (COWAT; Benton et al., 1994)
The COWAT is a test of verbal fluency. Participants were asked to generate as many words as possible that start with a specific letter (F, A, and S) in one minute. Proper nouns, words with different endings (e.g., -s, -ing, -ed) and repeated words were excluded. Aged-normed T-scores were used from the Delis-Kaplan Executive Functioning System assessment battery (D-KEFS; Delis et al., 2001). Higher scores indicated better verbal fluency.
Stroop Colour-Word Test (Golden & Freshwater, 1978)
The Stroop Colour-Word Test is a commonly-used measure of inhibitory control, cognitive flexibility, and selective attention (Strauss et al., 2006). The Word trial, Colour trial, and Colour-Word trial were administered, with each phase requiring participants to answer as many items as possible within a 45-second interval. The Word trial requires participants to read aloud colour words (RED, BLUE, GREEN) printed in black ink. The Colour trial requires participants to name the printed ink colour of non-word stimuli (XXXX). The Colour-Word trial requires participants to name the ink colour of a colour-incongruent word (e.g., correctly naming the ink colour red when the word says “BLUE”). Aged-normed T-scores for the colour-word trial were used in this study (Golden & Freshwater, 1978).
Cognitive Restructuring (CR) Task
Participants completed a standardised cognitive restructuring task (Johnco et al., 2013b, 2014a, 2015a). Younger and older adults used the same cognitive restructuring form from the Ageing Wisely CBT program (called the ‘REPLACE Unhelpful Thoughts’ technique; Wuthrich, 2009). Children used a parallel cognitive restructuring form, from the Cool Kids CBT program (Rapee et al., 2019). The format of the cognitive restructuring forms for all age groups included ‘situation – thought – feeling and subjective units of distress (SUDs) rating (0-100 for adults/older adults, and 0–10 for children which was later converted to align across age groups)’, followed by prompting questions to help generate disconfirmatory evidence (e.g., “What happened when I was worried before?”). Finally, participants were prompted to generate a realistic thought based on the evidence generated, and to re-rate their SUDs.
For all age-groups, the CR task included an orientation phase, learning phase, practice phase and test phase. The orientation phase consisted of didactic instruction on the link between thoughts and feelings, including examples of how different thoughts can create different emotional reactions in the same situation. The learning phase involved demonstration of the procedure to use a structured CR form. For younger and older adults, these phases were facilitated by the experimenter. For children, the orientation and learning phases were administered using standardised video instructions from the internet-delivered version of the Cool Kids program, (i.e., Cool Kids Online; McLellan et al., 2015) to teach and demonstrate cognitive restructuring skills (called ‘Detective Thinking’). Children completed the practice and test phases using paper-based versions of the cognitive restructuring form.
During the practice phase, participants completed the cognitive restructuring form with support from the experimenter to challenge and modify a recent unhelpful or irrational thought. In the testing phase, participants were assisted by the experimenter to identify a target situation and thought, and completed the remainder of the form independently. This task was administered by graduate-level Clinical Psychology students and took approximately 60-minutes. The quality of cognitive restructuring skills was assessed by an independent Clinical Psychologist or graduate-level Clinical Psychology student according to previously developed scoring criteria (see Table 2; Johnco et al., 2013a), with higher scores reflecting better quality cognitive restructuring skills. Inter-rater reliability coding was conducted on a random 20% of forms, with excellent inter-rater reliability among children (ICC = 0.73), younger adults (ICC = 0.96) and older adults (ICC = 0.86). Efficacy of cognitive restructuring was assessed by examining reduction in SUDS pre-post cognitive restructuring, with higher scores reflecting greater reduction in distress.
Table 2
Cognitive Restructuring Task Quality Scoring (from Johnco et al., 2013b)
Situation | |
|---|---|
0 | Missing of irrelevant responses that fail to specify the situation or context in which the distressing emotion is experienced (e.g. “I feel upset”). |
1 | Marginal descriptions of the problem situation, or descriptions that are confounded with one’s automatic thoughts (e.g. “My neighbour purposely ignored me when they were watering the garden”). |
2 | Reasonably complete and “objective” descriptions of a specific situation triggering a negative emotion (e.g. “My husband is three hours late returning from the meeting without phoning me in advance”). |
Target Cognition | |
0 | Missing responses, or responses that are merely restatements of problem situation or emotional reactions (e.g. “He’s late again. I feel so lonely”). |
1 | Vague interpretations, or rhetorical questions that obscure the dysfunctional belief that has been activated (e.g. “Putting myself down”, “Why does this keep happening to me?”). |
2 | Clear identification of the mental imagery or stream of consciousness that sustains the specified emotion (e.g. “This just proves what a failure I am” “You can never really trust a man”). |
Emotion | |
0 | Missing responses, or responses that specify automatic thoughts rather than feelings (e.g. “I can’t take it anymore”). |
1 | Vague emotional descriptions (e.g. “bad” “lousy”), or specific emotions (e.g. “anxious” “sad”) unaccompanied by ratings of intensity of feeling. |
2 | Accurate identification and ratings of specific feelings uncontaminated by automatic thoughts (e.g. “guilty, 75”). |
Evidence Against the Thought | |
0 | Evidence is irrelevant to the automatic thought, missing, or is evidence that supports the thought (e.g. "last week I forgot to pay the electricity bill too”). |
1 | Weak attempt to challenge the automatic thought, or only 1 or 2 pieces of evidence generated that would be unlikely to reduce negative affect. |
2 | Reasonable attempt to challenge that negative thought. Several pieces of evidence generated that question the validity of the thought, and would reasonably reduce negative affect. |
3 | Strong attempt to challenge the negative thought. At least 3 pieces of strong evidence which would directly challenge the validity of the thought and would result in a reduction in negative affect. |
Replacement Thought/Rational Response | |
0 | Missing or inappropriate responses that fail to challenge the automatic thoughts identified (e.g. “I shouldn’t feel this way”). |
1 | Problem solving statements or problem solving responses with no direct challenging of the cognitions (e.g. in response to challenging a cognition about having no-one to talk to at a social function - “try to avoid future situations like this”, or challenging a cognition about having a heart attack - “be brave, drink water and pray”) |
2 | Weak attempts to dispute or disprove automatic thoughts, or responses that specify no clear adaptive perspective or behaviour (e.g. “Maybe it won’t happen”, “Hang in there”, “You’re just catastrophizing”.) |
3 | Realistic attempts to seek evidence that disputes the validity of automatic thoughts, or developing an adaptive alternative interpretation of the situation (e.g. “Just because my child got a divorce doesn’t mean I’m a failure as a parent”. “His behaviour stems from his alcoholism, and I can’t take the blame for that”.) |
Outcome | |
0 | Missing response or vague statement of a different feeling, other than those identified in Emotions column (e.g. “Better”). |
1 | Specification of emotional outcome, but unaccompanied by rating of intensity (e.g. “Still somewhat fearful”.) |
2 | Clear specification of previous emotion, with rating of its present level of intensity (e.g. Fearful, 25). |
Procedure
Ethics Approval for each of these three samples was granted by the Macquarie University Human Research Ethics Committee. Younger and older adults provided written consent. Parents and guardians provided written consent for children, with children providing written assent. All measures were completed in-person with older adults, and over zoom with younger adults due to COVID-19 lockdowns. For children, the YODA was completed by parents via Redcap prior to the in-person session, and the experimental session was completed in-person. The experimental session took approximately 2–3 h to complete neuropsychological testing, then cognitive restructuring tasks, including breaks. Participation was voluntary and no reimbursements were given. Eligible students in the younger adult sample were awarded course credit for participation. This study was not preregistered.
Data Analysis
Logistic regression was used to examine age differences in dichotomously coded cognitive restructuring quality (i.e., good vs. poor skills, defined based on total score of quality ratings for the Evidence and Replacement Thought sections of ≥ 4/6 reflecting reasonable or strong quality responses). ANOVA was used to assess age differences in continuous measures of cognitive restructuring quality and efficacy (i.e., SUDS reduction) scores, with post-hoc comparisons conducted using sidak-adjusted alpha levels. Bivariate Pearson’s correlations were conducted between cognitive restructuring scores and neuropsychological measures of cognitive flexibility for the total sample, and separated by age group. The relative contribution of age and cognitive flexibility, and the potential moderating role of age in the relationship between cognitive flexibility and cognitive restructuring was examined using regression analyses, including age group, neuropsychological measures of cognitive flexibility (WCST-PE, COWAT, and Stroop) and the three interaction terms between age and each cognitive flexibility measure as predictors. Logistic regression with the same predictors was used to predict categorical outcomes (good vs. poor) for cognitive restructuring skills. To examine potential mediation effects, multicategorical mediation was used. Given that there were three categorical age groups included in this study, standard mediation analyses are not appropriate. Multicategorical mediation analyses were performed using the PROCESS macro in SPSS (Hayes, 2018) with 5000 bootstrap samples given the ability to simultaneously examine indirect effects for each pairwise combination of age group within the same analytical framework.
Results
Preliminary Analyses
Age differences in demographic and clinical characteristics are summarised in Table 1. There were no significant demographic differences between age groups. Younger adults and older adults were significantly more likely to have a comorbid diagnosis, and a comorbid affective disorder compared to children, but there were no significant age-differences in the presence of a comorbid anxiety disorder. There was no significant difference in primary disorder severity (ADIS CSR) between younger adults and older adults (M = 5.81, SD = 0.78 vs. 5.98, SD = 1.03, respectively; F(1,78) = 0.60, p =.442). Severity was rated on a different scale for children (0–4 compared to 0–8 for adults and older adults), however when using standardized z-scores of disorder severity, there were also no significant differences between age-groups (F(2,113) = 0.30, p =.742). There were no significant sex differences in cognitive restructuring skills or neuropsychological test performance between males and females (all p’s > 0.05). Although there was a significant overall group differences in SUDs severity at the start of the cognitive restructuring task (Molder adult = 80.19, SD = 14.60; Myounger adult = 73.03, SD = 13.25; Mchild = 71.71, SD = 19.36; F(2,113) = 3.43, p =.036), post-hoc tests showed no significant pair-wise group differences (all p’s > 0.05).
Age Differences in Cognitive Restructuring and Neuropsychological Skills
Age differences in cognitive restructuring and cognitive flexibility skills are summarised in Table 3. Looking at categorical classification of good vs. poor cognitive restructuring skills, younger adults were significantly more likely to have good quality CR skills compared to older adults (OR = 6.16, B = 1.82, SE = 0.61, χ2(1) = 8.91, p =.003). There was no significant difference between older adults and children (OR = 2.54, B = 0.93, SE = 0.49, χ2(1) = 3.70, p =.054), or between younger adults and children (OR = 2.42, B = 0.89, SE = 0.66, χ2(1) = 1.80, p =.180). Results were similar using continuous measures of cognitive restructuring quality, with younger adults showing significantly better quality cognitive restructuring skills compared to both children and older adults (both p’s < 0.001), although there was no significant difference between children and older adults (p =.965). Older adults showed smaller SUDS reduction compared to both younger adults and children (p <.001 and p =.008, respectively), with no significant difference between these age groups (p =.399).
Table 3
Age differences in cognitive restructuring skills and neuropsychological measures of cognitive flexibility
Child (N = 35) | Younger Adult (N = 32) | Older Adult (N = 47) | ||
|---|---|---|---|---|
n (%) | n (%) | n (%) | χ2 | |
Good Quality Cognitive restructuring skilla | 26 (74.3%) | 28 (87.5%)1 | 25 (53.2%)1 | 11.63** |
M (SD) | M (SD) | M (SD) | F | |
Cognitive restructuring | ||||
Quality | 11.66 (1.47)1 | 13.22 (1.18)1,2 | 11.51 (1.80)2 | 13.04*** |
Efficacy | 49.57 (23.50)1 | 42.56 (14.23)2 | 28.38 (20.65)1,2 | 11.95*** |
Cognitive Flexibility | ||||
WCST-PE | 58.43 (7.77)1 | 55.94 (10.08) | 51.66 (11.58)1 | 4.70* |
Stroop Colour-Word | 46.54 (8.92) | 51.22 (7.70) | 47.11 (10.08) | 2.67 |
COWAT | 58.06 (11.23) | 52.22 (11.72) | 52.87 (11.28) | 2.81 |
Older adults had significantly more WCST perseverative errors (WCST-PE) compared to children (p =.010), but not younger adults (p =.190). However, there were no other significant age-differences in cognitive flexibility on the Stroop or COWAT.
Relationship between Cognitive Flexibility and Cognitive Restructuring Skills
Bivariate correlations between cognitive flexibility and cognitive restructuring skills in the total sample and separated by age group are included in Table 4. There was a significant relationship between lower WCST perseverative errors and better CR quality (but not efficacy) in younger adults and older adults. Although there was a small to moderate correlation between Stroop performance and CR quality in children (r =.31, p =.071) this was not statistically significant. There was no evidence of a relationship between cognitive flexibility and cognitive restructuring skills in children.
Table 4
Correlations between cognitive flexibility and cognitive restructuring in the total sample, and separated by age group
CR Quality | CR Efficacy | WCST-PE | Stroop | COWAT | ||
|---|---|---|---|---|---|---|
Total sample | CR Efficacy | 0.22* | - | |||
WCST-PE | 0.27** | .23* | - | |||
Stroop | 0.19* | 0.07 | 0.03 | - | ||
COWAT | 0.05 | 0.15 | 0.25** | 0.31*** | - | |
Children | CR Efficacy | 0.06 | - | |||
WCST-PE | 0.07 | 0.07 | - | |||
Stroop | 0.31 | − 0.07 | 0.14 | - | ||
COWAT | 0.01 | 0.00 | 0.07 | 0.41* | - | |
Adults | CR Efficacy | 0.41* | - | |||
WCST-PE | 0.44* | 0.33 | - | |||
Stroop | 0.13 | − 0.13 | 0.23 | - | ||
COWAT | − 0.07 | 0.01 | 0.21 | 0.34 | - | |
Older adults | CR Efficacy | 0.23 | - | |||
WCST-PE | 0.29* | 0.10 | - | |||
Stroop | 0.01 | 0.23 | 0.24 | - | ||
COWAT | 0.24 | 0.23 | 0.30* | .35* | - |
Multivariate regression models examined whether there were age differences in cognitive restructuring skill after accounting for differences in cognitive flexibility skills, and whether the relationship between cognitive flexibility and cognitive restructuring skill acquisition differed between the age groups (i.e., moderation by age). Age group, WCST-PE, COWAT, Stroop and the interaction between age group and cognitive flexibility measures (WCST-PE, COWAT, Stroop) were entered into the model predicting cognitive restructuring quality, F(14,114) = 3.12, p <.001, ηp2 = 0.30. There was a significant main effect of WCST-PE (F(1,102) = 4.48, p =.037, ηp2 = 0.04), while the main effects of age group, COWAT, Stroop and the age group by cognitive flexibility interactions were not significant (all p’s > 0.05), suggesting that the relationship between cognitive flexibility and cognitive restructuring quality did not differ between the age groups.
In the regression model predicting cognitive restructuring efficacy (SUDS reduction), there were no significant main effects of agegroup, cognitive flexibility, or cognitive flexibility by agegroup interactions (all p’s > 0.05).
Cognitive Flexibility as a Mediator of age Differences in Cognitive Restructuring Skills
Given WCST-PE was the only index of cognitive flexibility to show consistent relationships with cognitive restructuring skills, this was examined as a potential mediator (see Fig. 1). Multicategorical mediation analyses showed that WCST-PE mediated the age-difference in CR skills between older adults and children (indirect effect = −.17, SE =.09, 95%CI = −.37 to −.03), and there was suggestion that it also mediated the difference in cognitive restructuring skills between older adults and younger adults given the pattern and magnitude of effects, and that the asymmetric distribution of the indirect effect confidence interval only marginally crosses 0 (indirect effect = −.10, SE =.08, t = 4.12, p <.001, 95%CI = −.28 to.01). However, there was less evidence that it mediated age differences in cognitive restructuring skills between younger adults and children given the indirect path crossed 0 (indirect effect = −.06, SE =.06, 95%CI = −.22 to.04), and the direct relationship (c’) was strengthened with the inclusion of WCST perseverative errors as a mediator.
Fig. 1
Multicategorical mediation analyses ofperseveration as a mediator of age-differences in cognitive restructuring ability. *p < .05, **p < .01, ***p < .001. WCST = Wisconsin Card Sorting Test
×
![]()
Discussion
This study examined whether there were developmental differences in cognitive restructuring skill acquisition between children, younger adults and older adults, and whether these age differences were moderated and/or mediated by cognitive flexibility skills. With only one hour of instruction, results indicated that between 53 and 87% of participants were able to effectively learn cognitive restructuring, with the remaining demonstrating suboptimal performance. Results suggest that cognitive flexibility is important when learning cognitive restructuring skills, particularly for younger and older adults, although perhaps less-so for children. Specifically, those who show greater tendency to perseverate, also tended to show more difficulty learning cognitive restructuring. Although univariate relationships suggested that older adults and children may show, on average, more difficulty learning cognitive restructuring skills than younger adults, this was explained by age-differences in cognitive flexibility skills – specifically, perseveration. After accounting for age differences in perseveration, there were no age-differences in ability to learn cognitive restructuring. In addition, perseveration mediated the age differences in cognitive restructuring skills between older adults and both younger adults and children, suggested that age-differences in cognitive restructuring quality are explained, at least in part, by differences in perseveration.
Overall, these results suggest that perseveration may be an important cognitive skill required for initial learning of cognitive restructuring, particularly among younger adults and older adults. However there was limited evidence to suggest that cognitive flexibility ability is influential in the ability to learn cognitive restructuring among children. This aligns with previous research showing cognitive flexibility skills are related to the ability to learn cognitive restructuring in both younger and older adults (Holder et al., 2021; Johnco et al., 2013a, 2014a, 2015a; Mohlman, 2013), and presents new knowledge in children. Although multivariate analyses suggest that the relationship between cognitive flexibility and cognitive restructuring skills did not differ between age groups, bivariate analyses showed that there was little association between measures of cognitive flexibility and cognitive restructuring skills in children. There are several reasons that cognitive flexibility may be less influential during childhood, compared to adulthood and older adulthood. It is possible that other types of cognitive ability are more influential, for example abstract thinking, metacognitive awareness, self-reflection and insight ability, all of which undergo significant maturation across childhood and adolescence (Dumontheil, 2014; Weil et al., 2013). In addition, while adult cognitions are often internal mental processes, children with anxiety often seek (and receive) reassurance from parents. It is possible that parental reassurance may be remembered and provided during cognitive restructuring tasks regardless of their cognitive ability to generate this evidence independently, or that cognitive challenging may be more of a dyadic process in this age group. While evidence suggests deficits in executive functions in younger and older adults with anxiety disorders (Yochim et al., 2013; Zainal & Newman, 2022), there is less evidence suggesting comparable deficits in executive functioning, including inhibitory control, among children with anxiety (Rabner et al., 2024). As such, there may be differential cognitive sequalae of anxiety in children compared to adults. Finally, while adults volunteered to participate in the studies and may consequently demonstrate greater willingness to self-reflect on their cognitions, children were volunteered by their parents and may have more variability in their willingness and cognitive avoidance.
This is the first study to examine age-differences in cognitive restructuring skills across the lifespan, as well as the potential mechanisms underlying the ability to use this skill effectively. It also extends previous research by using and separating multiple neuropsychological tests of cognitive flexibility, which provides a more comprehensive assessment of cognitive skills, and is superior to questionnaire-based assessment of cognitive ability (Howlett et al., 2022; Johnco et al., 2014b). However there are several limitations to note. Firstly, this study examined the role of cognitive flexibility on the ability to initially learn cognitive restructuring during a brief, one-hour, session. While naturally occurring cognitive abilities may affect initial cognitive restructuring skill acquisition, skill performance is likely to improve with repetition and training. Results may not extend to the ability to learn cognitive restructuring across the course of treatment, where clients practice and receive feedback from therapists to improve their skill utilisation. Indeed, Johnco et al. (2014a) found that pre-treatment cognitive flexibility skills did not predict the ability to technically use cognitive restructuring by the end of treatment, although it continued to predict therapist perceptions about how well clients were able to use cognitive restructuring. It may be that clients with executive dysfunction can learn to provide technically correct responses with practice, but that the ability to genuinely and affectively shift their ways of thinking continues to be more challenging for those with greater perseveration. Second, the adult and child samples were not treatment-seeking samples, and not all participants were treatment naïve. As such, results warrant replication. Given the variability in research on the relationship between executive functioning and clinical outcomes, future research should also examine the role of specific cognitive abilities on other CBT skills (e.g., problem solving, exposure therapy). Finally, mediation analyses used cross-sectional data which precludes the ability to establish temporal precedence of variables, and to draw causal conclusions. As such, results can only infer mediation and should be interpreted with caution.
There is an increasing interest in the potential for individual client characteristics to be used in tailoring of treatment. It is possible that cognitive flexibility, and specifically perseverative tendencies, may represent a candidate target for future research on treatment personalisation that examines the impact of emphasising or de-emphasising cognitive restructuring skills in treatment. However it is important to note that Johnco et al. (2014a) found that pre-treatment cognitive flexibility skills did not predict the ability to learn cognitive restructuring by the end of treatment and did not predict clinical outcomes, suggesting that omission of cognitive restructuring for all older adults with poor cognitive flexibility may not be necessary or appropriate. It is possible that cognitive flexibility is a relevant clinical characteristic to guide treatment tailoring, but may not necessarily prevent positive treatment outcomes.
Cognitive restructuring is a core cognitive skill in CBT, however research has primarily focused on its efficacy rather than the mechanisms underpinning the ability to learn it at different ages. These findings suggest that cognitive flexibility (specifically, the ability not to perseverate), is important during the initial stages of learning cognitive restructuring, particularly for adults and older adults. In contrast, there was minimal evidence to suggest that cognitive flexibility plays a role in learning cognitive restructuring among children. After considering individual differences in perseveration, there was no evidence that older adults show greater difficulty learning cognitive restructuring compared to any other agegroup. Individual differences in perseveration, rather than chronological age, may be a more important factor to consider in treatment modifications for adults and older adults.
Declarations
Competing Interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.