دانلود مقاله ISI انگلیسی شماره 38689
ترجمه فارسی عنوان مقاله

روابط بین توانایی های سرگردانی ذهن و کنترل توجه در بزرگسالان و نوجوانان

عنوان انگلیسی
Relationships between mind-wandering and attentional control abilities in young adults and adolescents
کد مقاله سال انتشار تعداد صفحات مقاله انگلیسی
38689 2015 12 صفحه PDF
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Acta Psychologica, Volume 148, May 2014, Pages 25–36

ترجمه کلمات کلیدی
سرگردانی ذهن - حواس پرتی - توانایی کنترل توجه - حافظه کاری - نوجوانی
کلمات کلیدی انگلیسی
Mind-wandering; Distractions; Attentional control abilities; Working memory; Adolescence
پیش نمایش مقاله
پیش نمایش مقاله  روابط بین توانایی های سرگردانی ذهن و کنترل توجه در بزرگسالان و نوجوانان

چکیده انگلیسی

Abstract Recent findings suggest that mind-wandering—the occurrence of thoughts that are both stimulus-independent and task-unrelated—corresponds to temporary failures in attentional control processes involved in maintaining constant task-focused attention. Studies supporting this proposal are, however, limited by a possible confound between mind-wandering episodes and other kinds of conscious experiences, such as external distractions (i.e., interoceptive sensations and exteroceptive perceptions). In the present study, we addressed this issue by examining, in adolescents and young adults, the relations between tasks measuring attentional control abilities and a measure of mind-wandering that is distinct from external distractions. We observed (1) that adolescents experienced more frequent external distractions, but not more mind-wandering, than young adults during the Sustained Attention to Response Task (SART) and (2) that, in young adults, the influence of external distractions on SART performance was fully accounted for by attentional control abilities, whereas mind-wandering was associated with decreases in SART performance above and beyond what was explained by attentional control abilities. These results show that mind-wandering cannot be entirely reduced to failures in the ability to maintain one's attention focused on task, and suggest that external distractions rather than mind-wandering are due to attentional control failures.

مقدمه انگلیسی

1. Introduction Mind-wandering refers to the occurrence of stimulus-independent and task-unrelated thoughts (Singer, 1993, Smallwood and Schooler, 2006, Stawarczyk, Majerus, Maj, Van der Linden and D'Argembeau, 2011 and Stawarczyk, Majerus, Maquet and D'Argembeau, 2011). While reading a book, for instance, it sometimes happens that our mind drifts away from the text and focuses instead on internal thoughts whose content is unrelated to the present situation, like memories or prospective thoughts. Recent research has revealed that mind-wandering represents a substantial part of our daily thinking time (i.e., from 20 to 50%; Kane, Brown, et al., 2007, Killingsworth and Gilbert, 2010 and Song and Wang, 2012), an important part of which is directed towards planning and preparing for future events (Baird et al., 2011, Song and Wang, 2012, Stawarczyk, Cassol and D'Argembeau, 2013 and Stawarczyk, Majerus, Maj, Van der Linden and D'Argembeau, 2011). Mind-wandering is commonly associated with decreased performance on the task performed at the moment of its occurrence. For instance, mind-wandering during reading has been consistently associated with decreased text comprehension (McVay and Kane, 2012b, Smallwood, 2011 and Unsworth and McMillan, 2013), and the occurrence of mind-wandering during go/no-go tasks has been related to more variable reaction times (RTs) to the go stimuli and an increased rate of errors to the no-go stimuli (Cheyne et al., 2009, McVay and Kane, 2009, McVay and Kane, 2012a, Stawarczyk, Cassol and D'Argembeau, 2013 and Stawarczyk, Majerus, Maj, Van der Linden and D'Argembeau, 2011). Another typical finding is that the frequency of mind-wandering is generally high during relatively low-demanding and easy tasks, and gradually decreases with increasing difficulty and task demands (McKiernan et al., 2006 and Smallwood and Schooler, 2006). Together, these findings suggest a close relationship between the occurrence of mind-wandering and attentional control processes (i.e., the domain general ability to maintain one's attention focused on a specific aspect of the environment). The nature of this relationship still remains debated, however. Two main theories have been proposed to account for the relationship between mind-wandering and attentional processes. On the one hand, the perceptual decoupling theory of mind-wandering (Schooler et al., 2011, Smallwood, 2010 and Smallwood and Schooler, 2006) suggests that mind-wandering results from a redirection of attentional resources from the task at hand to the processing and maintenance of internal thoughts (Levinson, Smallwood, & Davidson, 2012). In this proposal, mind-wandering is a resource consuming phenomenon that is more frequent during easier tasks because a larger amount of cognitive resources are available to support internal thoughts in comparison to more difficult tasks (Smallwood & Schooler, 2006). On the other hand, for the control failure theory (McVay and Kane, 2010b and McVay and Kane, 2010a), mind-wandering does not recruit attentional resources; instead, the occurrence of mind-wandering would reflect a temporary breakdown in attentional control processes that are involved in maintaining task-focused attention. According to this view, individuals with low attentional control abilities are more likely to experience mind-wandering because they are less efficient in maintaining their attention on the ongoing task (McVay & Kane, 2009). This theory entails that the lower frequency of mind-wandering during more difficult tasks is due to the fact that higher task demands lead to a continuous recruitment of attentional control processes during task completion preventing the occurrence of mind-wandering (McVay & Kane, 2010b). The control failure theory of mind-wandering is indirectly supported by the finding that individuals known to have decreased attentional control abilities, like people under the influence of alcohol (Finnigan et al., 2007 and Sayette et al., 2009) or college students who received a diagnosis of attention deficit–hyperactivity disorder during childhood (Hines and Shaw, 1993 and Shaw and Giambra, 1993), report experiencing a higher frequency of mind-wandering. Furthermore, several studies have found a negative relationship between working memory capacity, as assessed with complex span tasks (Conway et al., 2005 and Redick et al., 2012), and the frequency of mind-wandering sampled during both laboratory tasks (McVay and Kane, 2009, McVay and Kane, 2012a, McVay and Kane, 2012b and Unsworth and McMillan, 2013) and daily life activities (Kane, Brown, et al., 2007) that are challenging in terms of attentional demands. Working memory capacity actually measures a domain general attentional control ability that corresponds to the maintenance of goal-relevant information in the focus of attention (Engle and Kane, 2004 and Kane, Conway, Hambrick and Engle, 2007). Individuals with high working memory capacity might thus experience less mind-wandering because they possess better attentional control abilities, which allow them to stay focused on demanding tasks to a larger extent than individuals with low working memory capacity (McVay and Kane, 2010b and McVay and Kane, 2010a). A possible limitation of the studies that showed a negative relationship between mind-wandering frequency and working memory capacity is that mind-wandering was operationalized as the occurrence of any task-unrelated thoughts, without consideration of whether or not these thoughts were also stimulus-independent. Indeed, a category of conscious experiences labeled as “current state of being” (defined as “thoughts about being sleepy, hungry, bored, or any other current state”) was considered as mind-wandering episodes in these studies (McVay and Kane, 2009, McVay and Kane, 2010b, McVay and Kane, 2012a, McVay and Kane, 2012b and Unsworth and McMillan, 2013). This might be an important issue to take into consideration because a central aspect of the definition of mind-wandering is that the content of these thoughts is unrelated to current sensory input (Schooler et al., 2011 and Smallwood et al., 2012). It has been suggested that distractions by directly perceived stimuli might involve different cognitive processes than distractions by internally generated thoughts that do not have a direct referent in the current environment (Friedman and Miyake, 2004, Gilbert et al., 2007 and Lustig et al., 2001). For instance, using latent variable analyses, Friedman and Miyake (2004) showed that tasks in which distractor stimuli are visually presented together with the target stimuli load on a different latent variable than tasks involving the resistance to mental interference resulting from information presented prior to the target stimuli. Furthermore, these two kinds of tasks correlated with different measures of individual differences: the former latent variable was associated with the occurrence of cognitive failures in daily life, while the latter was associated with a general tendency to experience intrusive thoughts (see also Verwoerd et al., 2009 and Verwoerd et al., 2011). Analyses of task performance have shown that mind-wandering and distractions by sensory input (referred to as “external distractions”) are both associated with commission errors and more variable RTs during go/no-go tasks (Stawarczyk, Majerus, Maj, Van der Linden and D'Argembeau, 2011 and Stawarczyk, Majerus, Maquet and D'Argembeau, 2011). However, neuroimaging evidence suggests that these two types of experiences are not equivalent: although both are associated with activity in the default mode network, mind-wandering induces significantly more activation in this network compared to external distractions (Kucyi et al., 2013 and Stawarczyk, Majerus, Maquet and D'Argembeau, 2011). External distractions occur when individuals stop being fully focused on a task because of thoughts about exteroceptive perceptions or interoceptive sensations that are unrelated to this task (e.g., being distracted from reading a book because of a sudden phone ring or because one begins to feel hungry), which corresponds to the above mentioned “current state of being” experience. Intriguingly, in the studies that conceptualized mind-wandering as task-unrelated thoughts without taking stimulus-independence into account, the “current state of being” experiences represented around 50% of mind-wandering episodes1 (e.g., McVay and Kane, 2009 and McVay and Kane, 2012a), and some indirect evidence suggests that these two categories of experiences may be differently related to working memory capacity. Indeed, a recent study has shown that the negative correlation between mind-wandering frequency and working memory capacity is much less consistent when the “current state of being” experiences are not included in the analyses; past-oriented mind-wandering was unrelated to working memory capacity and a significant negative correlation between future-oriented mind-wandering and working memory capacity was only found in one of two samples of participants (McVay, Unsworth, McMillan, & Kane, 2013). This latter study did not examine how “current state of being” experiences are specifically associated with working memory capacity, however, and it thus remains unclear how external distractions and mind-wandering relate to working memory capacity (and attentional control abilities in general) when they are clearly distinguished from one another. From the current state of findings, we cannot dismiss the possibility that previously documented associations between mind-wandering and attentional control measures were actually attributable, at least partially, to the frequency of external distractions. In the present study, we sought to investigate this issue with the use of thought-probes that clearly distinguish mind-wandering from external distractions (Stawarczyk, Majerus, Maj, et al., 2011) during the Sustained Attention to Response Task (SART; Robertson, Manly, Andrade, Baddeley, & Yiend, 1997). To measure attentional control abilities, participants carried out a typical complex span tasks and the AX version of the continuous performance task (AX-CPT; Braver et al., 2001), a task that assesses both proactive and reactive attentional control abilities (Braver, 2012, Braver et al., 2007 and Iselin and DeCoster, 2009). McVay and Kane (2010b) have indeed suggested that these two forms of attentional control might be important in determining the frequency of mind-wandering. Proactive control reflects the sustained and anticipatory maintenance of goal-relevant information in order to enable optimal cognitive performance, which might be crucial for preventing the occurrence of mind-wandering. Reactive control, on the other hand, reflects a transient activation of goal-related information in response to a triggering stimulus and might be involved in the ability to suppress mind-wandering after its occurrence in order to get back on task (McVay & Kane, 2010b). Finally, we not only included young adults in this study, as in most previous studies of mind-wandering, but also adolescents. It has been shown that attentional control abilities are still developing during adolescence (De Luca et al., 2003, Fry and Hale, 1996, Iselin and DeCoster, 2009 and Siegel, 1994), making this age group an adequate canditade to examine whether lower and more variable attentional control abilities come along with a higher rate of mind-wandering and external distractions. To the best of our knowledge, no study to date has contrasted the frequency of mind-wandering in adolescence and young adulthood. Our hypotheses were the following: if the occurrence of mind-wandering reflects temporary failures in attentional control abilities (McVay and Kane, 2010b and McVay and Kane, 2010a) rather than a specific state of attention in which attentional resources are directed to the processing of internal thoughts (Schooler et al., 2011, Smallwood, 2010 and Smallwood and Schooler, 2006), then (1) adolescents should experience mind-wandering episodes during the SART to a larger extent than young adults, as the former typically show lower attentional control abilities than the latter (Iselin & DeCoster, 2009); (2) the frequency of mind-wandering should be negatively related to working memory capacity, as well as proactive and reactive attentional control abilities, when mind-wandering is clearly distinguished from external distractions; and (3) the effect of mind-wandering on SART performance should overlap for the most part with the effect of attentional control abilities in multiple regression models, and mind-wandering should therefore not remain an independent predictor of SART performance once the measures of attentional control abilities are taken into account.

نتیجه گیری انگلیسی

. Results 3.1. Group comparisons We first computed a series of Student's t-tests to examine whether young adults and adolescents differed in terms of SART performance, responses to the thought-probes, proactive and reactive attentional control, working memory capacity, and general fluid intelligence. Results of these analyses are shown in Table 1. Regarding SART performance, adolescents were slower and had more variable RTs to the non-target stimuli, and they also committed more errors to the target stimuli. As expected, adolescents showed lower and more variable performance on measures of proactive and reactive attentional control, working memory capacity, and general fluid intelligence. It is important to note, however, that the frequency of mind-wandering episodes reported during the SART was equivalent for young adults and adolescents. Young adults even reported more mind-wandering than adolescents on the DDFS. Regarding the other responses to the thought-probes, young adults reported being fully focused on task to a larger extent than adolescents, whereas adolescents reported more frequent external distractions; the two groups did not differ on the frequency of task-related interferences. Table 1. Comparisons between young adults and adolescents. Mean score (standard deviation) t(162) p Cohen's d Adolescents Young adults SART target accuracy (%) 50.28 (17.57) 68.41 (15.52) − 7.02 < .001 1.10 SART non-target RT (ms) 395 (45) 377 (46) 2.47 .01 .40 SART non-target CV 31.88 (8.94) 24.53 (7.50) 5.72 < .001 .90 % On-task reports 31.43 (20.32) 38.47 (21.67) − 2.14 .03 .34 % TRI reports 26.49 (16.33) 27.74 (15.01) − .51 .61 .08 % ED reports 24.85 (16.19) 16.63 (10.80) 3.86 < .001 .61 % MW reports 17.23 (12.96) 17.16 (16.76) .03 .98 .005 DDFS 40.29 (8.61) 42.94 (7.73) − 2.08 .04 .33 AX-CPT d′-proactive 2.28 (.79) 3.02 (.85) − 5.78 < .001 .91 AX-CPT d′-reactive 2.22 (.80) 3.09 (.69) − 7.39 < .001 1.18 WMC .85 (.09) .90 (.06) − 4.75 < .001 .67 Raven's matrices 20.10 (3.53) 23.18 (3.37) − 5.72 < .001 .90 Note: RT = reaction time; CV = coefficient of variation of RTs; TRI = task-related interference; ED = external distraction; MW = mind-wandering; DDFS = Daydreaming Frequency Scale; WMC = working memory capacity. Table options Next, we examined whether mind-wandering and external distractions impaired SART performance to the same extent in young adults and adolescents. To do so, we analyzed the RTs (means and CVs) to the five last non-targets of each block (Kam et al., 2011, Seli, Cheyne and Smilek, 2013 and Stawarczyk, Majerus, Maj, Van der Linden and D'Argembeau, 2011) as a function of the responses given to the probes, using a series of 2 (group) × 4 (probe response) mixed-design Analyses of Variance (ANOVAs). We also used a 2 × 4 ANOVA to analyze the proportion of correct target responses within each block, as a function of the responses given to the probes. For each measure (means and CVs of RTs for the non-targets, and accuracy to the targets), a single average score was computed per participant for each of the four kinds of thought-probe responses. As shown in panel A of Fig. 1, the ANOVA for mean RTs revealed main effects of group [F(1,131) = 7.96; p = .006; ηp2 = .06] and probe response [F(3,393) = 3.37; p = .02; ηp2 = .03], but no significant interaction effect [F(3,393) = 1.49; p = .22; ηp2 = .01]. Planned comparisons revealed that participants were faster when they reported being fully focused on task compared to when they experienced mind-wandering and external distractions [F(1,131) = 8.72; p = .004; ηp2 = .06], and that mean RTs did not differ between the two latter kinds of probe responses [F(1,131) = .05; p = .82; ηp2 < .001]. As shown in panel B, the ANOVA performed on CVs also revealed main effects of group [F(1,131) = 34.03; p < .001; ηp2 = .21] and probe response [F(3,393) = 8.99; p < .001; ηp2 = .06], but no significant interaction effect [F(3,393) = 2.35; p = .07; ηp2 = .02]. Planned comparisons showed that RTs were more variable when participants reported mind-wandering and external distractions than when they reported being fully focused on task [F(1,131) = 20.69; p < .001; ηp2 = .14]; there was also a trend for RTs to be more variable preceding mind-wandering than external distraction reports [F(1,131) = 3.86; p = .051; ηp2 = .03]. Finally, as shown in panel C, the ANOVA performed on target accuracy revealed significant main effects of group [F(1,131) = 35.71; p < .001; ηp2 = .21] and probe response [F(3,393) = 32.15; p < .001; ηp2 = .20], and again no significant interaction effect [F(3,393) = 1.63; p = .18; ηp2 = .01]. Planned comparisons showed that participants committed less errors during blocks when they reported that they were fully focused on task than when they reported mind-wandering and external distractions [F(1,131) = 62.96; p < .001; ηp2 = .32], and target accuracy did not differ between the two latter kinds of reports [F(1,131) = 1.47; p = .23; ηp2 = .01]. SART performance according to the responses given to the thought-probes for the ... Fig. 1. SART performance according to the responses given to the thought-probes for the two groups of participants. Note: bars represent the standard error on the mean; RTs = reaction times; CVs = coefficients of variation of RTs; TRI = task-related interference; ED = external distraction; MW = mind-wandering. Figure options In sum, the results of these ANOVAs show that participants' performance on the SART was worse when they experienced mind-wandering and external distractions compared to when they were fully focused on task (see also; Stawarczyk, Majerus, Maj, et al., 2011). The degree to which performance was affected by mind-wandering and external distractions was equivalent for adolescents and young adults. 3.2. Correlation analyses Next, we performed correlation analyses to examine the relationships between the different variables. These analyses were performed separately for the two groups of participants. As shown in Table 2, the correlations between SART performance and thought-probe responses revealed that mind-wandering reports were associated with lower performance (more target errors and a larger variability of RTs for non-targets) in young adults, whereas reports of being fully focused on task showed the opposite pattern of associations (they were related to less RT variability and fewer errors to the target stimuli). External distraction reports were associated with more target errors. These results confirm that individual differences in SART performance are related to mind-wandering and external distraction frequency (Stawarczyk, Majerus, Maj, et al., 2011). The analyses for the adolescent group (see Table 3) revealed only two significant correlations, which showed that more frequent on-task reports were associated with fewer errors to the target stimuli and that more frequent mind-wandering reports were associated with a higher variability of RTs for the non-target stimuli. Interestingly, correlations between the DDFS and responses to the thoughts probes were nearly identical in the two groups: the frequency of mind-wandering in daily life was associated with more reports of mind-wandering during the SART and fewer reports of being focused on task, but was not significantly associated with task-related interferences and external distractions. These results suggest that adolescents properly followed the instructions regarding the thought-probes. As DDFS scores were unrelated to all the other measures under investigation here, they will not be analyzed further. Table 2. Correlation matrix for the young adult group. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 1. % on-task .87 2. % TRIs − .41 p < .001 .73 3. % EDs − .56 p < .001 .007 p = .95 .63 4. % MW − .56 p < .001 − .37 p = .001 .08 p = .47 .86 5. SART acc. .45 p < .001 .04 p = .70 − .28 p = .009 − .44 p < .001 .84 6. SART CV − .27 p = .01 − .01 p = .91 .09 p = .38 .31 p = .004 − .34 p = .001 .94 7. SART RTs − .12 p = .28 .12 p = .25 .06 p = .58 < .001 p = .99 .05 p = .63 .51 p < .001 .97 8. WMC .33 p = .002 − .09 p = .43 − .33 p = .002 − .14 p = .19 .41 p < .001 − .32 p = .003 − .15 p = .16 .67 9. Raven .22 p = .04 .005 p = .96 − .06 p = .57 − .25 p = .02 .32 p = .003 − .17 p = .11 − .09 p = .42 .39 p < .001 .72 10. d′ proact. .27 p = .01 − .04 p = .70 − .16 p = .13 − .21 p = .06 .38 p < .001 − .25 p = .02 − .22 p = .04 .29 p = .006 .23 p = .04 11. d′ react. .38 p < .001 − .09 p = .39 − .24 p = .02 − .26 p = .02 .53 p < .001 − .16 p = .15 − .07 p = .54 .29 p = .006 − .18 p = .10 .63 p < .001 12. Att. comp. .42 p < .001 − .08 p = .48 − .28 p = .009 − .30 p = .005 .58 p < .001 − .31 p = .003 − 19 p = .09 .70 p < .001 .63 p < .001 .76 p < .001 .74 p < .001 13. DDFS − .23 p = .03 − .08 p = .43 .10 p = .36 .31 p = .003 − .17 p = .12 .07 p = .51 − .19 p = .09 .03 p = .80 .004 p = .97 .04 p = .70 − .001 p = .95 .02 p = .83 .88 Note: Italicized values on the diagonal reflect Cronbach's alpha for each measure as a reliability estimate (when applicable); alphas were calculated over task blocks for the measures of SART performance as well as responses to the thought-probes and over items for the other variables; TRIs = task-related interferences; EDs = external distractions; MW = mind-wandering; SART acc. = accuracy to the target stimuli; SART CV = coefficients of variation of RTs for the non-target stimuli; SART RTs = mean RT for the non-target stimuli; WMC = working memory capacity; Att. comp. = Attentional composite z-score of AX-CPT, WMC and Raven's matrices performances; DDFS = Daydreaming Frequency Scale. Table options Table 3. Correlation matrix in the adolescent group. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 1. % on-task .72 2. % TRIs − .40 .78 p < .001 3. % EDs − .49 − .33 .79 p < .001 p = .003 4. % MW − .43 − .20 − .05 .75 p < .001 p = .07 p = .65 5. SART acc. .29 − .17 − .09 − .13 .84 p = .01 p = .14 p = .44 p = .27 6. SART CV − .03 − .05 − .19 .27 − .12 .93 p = .83 p = .67 p = .09 p = .02 p = .32 7. SART RTs .10 .007 − .13 − .005 .30 .34 .95 p = .38 p = .95 p = .25 p = .97 p = .009 p = .002 8. WMC .13 − .02 − .06 − .11 .31 .03 − .25 .79 p = .26 p = .86 p = .63 p = .35 p = .006 p = .79 p = .03 9. Raven − .09 .15 − .008 − .04 .13 .07 − .04 .33 .71 p = .45 p = .20 p = .94 p = .73 p = .25 p = .55 p = .73 p = .003 10. d′ proact. .10 .05 .002 − .22 .47 − .32 − .19 .38 .26 p = .38 p = .67 p = .98 p = .052 p < .001 p = .005 p = .09 p = .001 p = .02 11. d′ react. .04 .04 .03 − .17 .55 − .42 − .05 .36 .28 .59 p = .71 p = .70 p = .77 p = .15 p < .001 p < .001 p = .64 p = .001 p = .01 p < .001 12. Att. comp. .06 .08 − .01 − .18 .50 − .22 − .19 .71 .65 .77 .77 p = .58 p = .51 p = .93 p = .11 p < .001 p = .055 p = .11 p < .001 p < .001 p < .001 p < .001 13. DDFS − .27 .04 .08 .27 .14 .22 .08 .05 .15 − .12 − .009 − .004 .87 p = .02 p = .75 p = .48 p = .02 p = .21 p = .054 p = .49 p = .65 p = .21 p = .31 p = .43 p = .97 Note: Italicized values on the diagonal reflect Cronbach's alpha for each measure (when applicable) as a reliability estimate; alphas were calculated over task blocks for the measures of SART performance as well as responses to the thought-probes and over items for the other variables; TRIs = task-related interferences; EDs = external distractions; MW = mind-wandering; SART acc. = accuracy to the target stimuli; SART CV = coefficients of variation of RTs for the non-target stimuli; SART RTs = mean RT for the non-target stimuli; WMC = working memory capacity; Att. comp. = Attentional composite z-score of AX-CPT, WMC and Raven's matrices performances; DDFS = Daydreaming Frequency Scale. Table options Next, we examined the relationships between proactive and reactive attentional control abilities, working memory capacity, and fluid intelligence. Results were consistent across the two groups (see Table 2 and Table 3) and showed that all variables significantly correlated with each other, except that reactive attentional control was not significantly correlated with fluid intelligence in the adult group. Finally, we examined whether the responses to the thought-probes and the measures of SART performance that were related to these responses (i.e., CVs and target accuracy) were related to proactive and reactive attentional control abilities, working memory capacity, and fluid intelligence. Furthermore, given the intercorrelations between the different cognitive tasks, we also computed an attentional composite Z-score (combining proactive and reactive attentional control abilities, working memory capacity, and fluid intelligence) that is free of the measurement error associated with each single task (Conway et al., 2005). Results showed that none of the four kinds of thought-probe responses was related to the different measures of cognitive abilities in the adolescent group (see Table 3). In young adults (see Table 2), mind-wandering was related to lower reactive attentional control and fluid intelligence. Furthermore, external distractions were related to lower reactive attentional control and working memory capacity, and reports of being fully focused on task were related to better proactive and reactive attentional control, working memory capacity, and fluid intelligence. These three kinds of responses to the thought-probes were also significantly related to the attentional composite Z-score (positively for on-task reports and negatively for mind-wandering and external distraction reports). Regarding SART performance, accuracy to the target stimuli significantly correlated with all measures of cognitive abilities in both groups of participants, with the exception of fluid intelligence for adolescents. RTs variability correlated to proactive and reactive attentional control in adolescents, and with proactive attentional control and working memory capacity in young adults. RTs variability was also correlated with the attentional composite Z-score in young adults and this association was nearly significant in the adolescent group (p = .055). Together, these results partially support the prediction stemming from the control failure theory that the frequency of mind-wandering should be related to proactive and reactive attentional control abilities, as well as working memory capacity. In the young adult group, mind-wandering frequency was indeed significantly related to reactive attentional control, fluid intelligence, and the attentional composite Z-score. On the other hand, however, the measures of cognitive abilities were unrelated to the frequency of mind-wandering in the adolescent group. These results suggest that mind-wandering is more closely tied to attentional control abilities in young adulthood than in adolescence. 3.3. Variance partitioning analyses As the results of the correlational analyses partly supported the predictions of the control failure theory of mind-wandering, we further explored the association between mind-wandering frequency, external distractions, SART performance and attentional control abilities. We used variance partitioning methods (e.g., Chuah and Maybery, 1999, Cowan et al., 2005 and Unsworth et al., 2009) to examine the shared and unique contribution of mind-wandering, attentional abilities and external distractions to SART performance. Variance partitioning attempts to allocate the overall R2 of a particular criterion variable (here accuracy to the target stimuli and variability of RTs to the non-target stimuli of the SART) into portions that are shared and unique to a set of predictor variables (mind-wandering, external distractions, and the attentional composite Z-score for target accuracy; mind-wandering and the attentional composite Z-score for RTs variability). These portions of the overall R2 are obtained by carrying out a series of regression analyses from different combinations of the predictor variables ( Unsworth et al., 2009). First, Table 4 shows that 41% of the variance in target accuracy during the SART in young adults was accounted for by the three predictor variables. Furthermore, as shown in Fig. 2, mind-wandering remained a significant predictor of SART accuracy beyond and above what was explained by the attentional composite Z-score and external distractions [t(83) = − 3.40; p = .001], explaining an additional 7% of the variance. On the other hand, external distractions were not an independent performance of accuracy to the target stimuli, explaining only 1% of additional variance beyond the two other variables [t(83) = − 1.49; p = .14]. As recent findings ( Seli, Jonker, Cheyne, & Smilek, 2013) have demonstrated that SART accuracy is a more direct measure of attentional failures after controlling for mean RTs (which reflect the way speed–accuracy trade-offs are handled), we also computed similar regression analyses with mean RTs as an additional independent variable. The inclusion of this variable did not change the significance of the results regarding mind-wandering, external distractions, and the attentional composite Z-score. Table 4. R2 values for regression analyses predicting SART accuracy for various predictor variables in the young adult group. Predictor variables Adjusted R2 F p 1. MW, EDs, AC .41 20.73 < .001 2. MW, EDs .24 14.42 < .001 3. MW, AC .40 29.55 < .001 4. EDs, AC .33 22.47 < .001 5. MW .19 20.68 < .001 6. EDs .07 7.15 .009 7. AC .33 42.56 < .001 Note: MW = mind-wandering; EDs = external distractions; AC = Attentional composite z-score of AX-CPT, WMC and Raven's matrices performances. Table options Venn diagram displaying the variance in target accuracy during the SART ... Fig. 2. Venn diagram displaying the variance in target accuracy during the SART accounted for by mind-wandering frequency, external distractions frequency, and the attentional composite z-score for the adult group. Figure options Second, Table 5 shows that mind-wandering and the attentional composite Z-score accounted for 13% of the variance in RTs variability in the young adult group. As shown in Fig. 3, mind-wandering significantly explained an additional 4% of the variance beyond and above attentional abilities [t(84) = 2.19; p = .03]. It should be noted that external distractions were not included in this analysis because they were not significantly correlated with the variability of RTs during the SART (r = .09; p = .38). Adding this variable did not change the significance of the results regarding mind-wandering and the attentional composite Z-score. Table 5. R2 values for regression analyses predicting SART variability of RTs for various predictor variables in the young adult group. Predictor variables Adjusted R2 F p 1. MW, AC .13 7.23 .001 2. MW .08 8.73 .004 3. AC .09 9.22 .003 Note: MW = mind-wandering; AC = Attentional composite z-score of AX-CPT, WMC and Raven's matrices performances. Table options Venn diagram displaying the variance in RTs variability to the non-target during ... Fig. 3. Venn diagram displaying the variance in RTs variability to the non-target during the SART accounted for by mind-wandering frequency and the attentional composite z-score for the adult group. Figure options Finally, Table 6 shows that mind-wandering and the attentional composite Z-score accounted for 8% of the variance in RTs variability in the adolescent group. As shown in Fig. 4, mind-wandering significantly explained an additional 4% of the variance beyond attentional abilities [t(74) = 2.07; p = .04]. Similarly to the analyses performed in the young adult group, external distractions were not included in this analysis because they were not significantly correlated with the variability of RTs during the SART (r = − .19; p = .09). Again, adding this variable did not change the significance of the results regarding mind-wandering and the attentional composite Z-score. Table 6. R2 values for regression analyses predicting SART variability of RTs for various predictor variables in the adolescent group. Predictor variables Adjusted R2 F p 1. MW, AC .08 4.13 .02 2. MW .06 5.67 .02 3. AC .04 3.81 .05 Note: MW = mind-wandering; AC = Attentional composite z-score of AX-CPT, WMC and Raven's matrices performances. Table options Venn diagram displaying the variance in RTs variability to the non-target during ... Fig. 4. Venn diagram displaying the variance in RTs variability to the non-target during the SART accounted for by mind-wandering frequency and the attentional composite z-score for the adolescent group.