Table 4.

The relationship between sedentary behaviour and cognitive development

No. of participants (No. of studies)	Design	Quality assessment					Absolute effect	Quality
		Risk of bias	Inconsistency	Indirectness	Imprecision	Other
The range of mean ages at time of exposure measurement was ~0.5 to 4.4 years; the oldest age range at follow-up was 9 to 10 years. Data were collected cross-sectionally and up to 8 years of follow-up. Cognitive development indicators were: ADHD symptoms (assessed by checklists based on the DSM-IV); attentional problems (assessed by the BPI); attention span (assessed by the CTQ); classroom engagement (assessed by a Classroom Engagement Scale and an unpublished questionnaire); cognitive ability (assessed by the Imitation Sorting Task); cognitive development (assessed by BSID-II and BSID-III); cognitive inhibitory control (assessed by the Animal Stroop Task); executive function (assessed as a composite of cognitive inhibitory control and working memory capacity; the BASC-2; four tasks: grass/snow, whisper, backward digit span, tower); language development (total), auditory comprehension, expressive communication (assessed by ASQ, PLS-4, CELF-P2, CELF-4, CDI, K-ASQ, Thai CLAMS, medical diagnosis, and developmental assessment with Denver-II test); mathematical success (assessed as relative to the class distribution); mathematics, reading recognition, reading comprehension (assessed by the PIAT); number knowledge (assessed by NKT); receptive and total vocabulary (assessed by PPVT); short-term memory (assessed by the Memory for Digit Span of the WISC); speech disorders (assessed by the Chuturik test and Child Behavior Checklist by Achenbach, conversation with parents, and clinical examination); and working memory capacity (assessed using the Animal Stroop Task and K-ABC number recall test).
8927 (11)	Longitudinala	Serious risk of biasb	No serious inconsistency	No serious indirectness	No serious imprecision	None	Screen-based sedentary behaviours: Electronic media exposure (duration): 1/1 studies reported unfavourable associations [112] Other sedentary behaviours: Parents reading (frequency): 1/1 studies reported favourable associations [121] TV time (duration): 5/10 studies reported unfavourable associations [90, 92, 100, 120, 121] 4/10 studies reported null associations [88, 102, 113, 122] 1/10 studies reported mixed unfavourable, null, and favourable associations [119]	Very lowc
166 (1)	Case-controld	Serious risk of biase	No serious inconsistency	No serious indirectness	No serious imprecision	None	Screen-based sedentary behaviours: TV time: 1/1 studies reported unfavourable associations [116]	Very lowf
9330 (16)	Cross-sectionalg	Serious risk of biash	No serious inconsistency	No serious indirectness	No serious imprecision	None	Objectively measured sedentary time: Total sedentary time (accelerometer-derived): 1/1 studies reported null associations [104] Screen-based sedentary behaviours: Computer use (yes, no): 1/1 studies reported null associations [109] Mobile phone use (yes, no): 1/1 studies reported unfavourable associations [109] TV time (duration): 3/9 studies reported unfavourable associations [94, 108, 123] 4/9 studies reported null associations [90, 100, 114, 115, 121] 1/9 studies reported mixed unfavourable and null associations [118] Total media exposure (duration): 1/1 studies reported mixed null and unfavourable associations [124] Video games (duration): 1/1 studies reported null associations [107] Other sedentary behaviours: Reading with parents (duration, frequency): 1/3 studies reported null associations [110] 1/3 studies reported favourable associations [117] 1/3 studies reported mixed null and favourable associations [124] Screen time (duration): 1/1 studies reported unfavourable associations [111] Storytelling with parents (frequency): 2/2 studies reported mixed null and favourable associations [117, 124]	Very lowi

ADHD Attention-Deficit/Hyperactivity Disorder, ASQ Ages and Stages Questionnaire, BASC-2 Behavior Assessment System for Children, BSID-II and BSID-III Bayley Scales of Infant Development–second and third editions, BPI Behavioral Problems Index, CDI Communicative Development Inventory, CELF-P2 Clinical Evaluation of Language Fundamentals–Preschool, CELF-4 Clinical Evaluation of Language Fundamentals Fourth Edition, CLAMS Clinical Linguistic Auditory Milestone Scale, CTQ Child Temperament Questionnaire, DSM-IV Diagnostic and Statistical Manual of Mental Disorders–4, K-ABC Kaufman Assessment Battery for Children, K-ASQ Korean–Ages and Stages Questionnaire, NKT Number Knowledge Test, PIAT Peabody Individual Achievement Test, PLS-4 Preschool Language Scale–4, PPVT Peabody Picture Vocabulary Test, WISC Wechsler Intelligence Scale for Children

^aIncludes 11 longitudinal studies [88, 90, 92, 100, 102, 112, 113, 119–122] from 8 unique samples. Tomopoulos et al. [112] reported data from the Bellevue Project for Early Language, Literacy, and Education Success (BELLE); McKean et al. [121] reported data from the Early Language in Victoria Study (ELVS); Pagani et al. [90, 92] reported data from the Quebec Longitudinal Study of Child Development (QLSCD); Schmidt et al. [88] reported data from Project Viva; and Foster and Watkins [113], Christakis et al. [120] and Zimmerman and Christakis [119] reported data from the National Longitudinal Survey of Youth, Children, and Young Adults (NLSY-Child). Results are presented separately and participants are counted only once

^bSerious risk of bias. Questionable validity and reliability of television duration exposure measure in all studies [88, 90, 92, 100, 102, 112, 113, 119–122]; poor reliability of Attention Problems subscale of the Child Behavior Checklist (ɑ =0.59) [102]; possible reporting bias, because the relationship between TV exposure and BMI at age 3 yr was analyzed despite not being described in the methods section [88]; two studies had unexplained missing data (34% and 40% missing) and the pattern of nonresponse indicates the reason for missing data may have been related to the outcome of interest [112, 121]; data were reported incompletely for the relationship between TV exposure and reading achievement [90]; the methods section of one study indicated that bivariate analysis would be performed, but included variables and the results of the analysis were not reported [121]

^cThe quality of evidence from longitudinal studies was downgraded from “low” to “very low” because of a serious risk of bias that diminished the level of confidence in the observed effects

^dIncludes 1 case-control study [116]

^eSerious risk of bias. Exposure measure was described in poor detail; questionable validity and reliability of television duration exposure measure; the Denver II Scale is useful for detecting severe developmental problems but has been criticized as being unreliable for predicting less severe or specific problems; the regression model that predicted developmental delay from a composite of “age of onset of TV viewing” and “TV viewing >2 h/day” was not pre-specified in the methods, and composite variables were not combined in analyses with other outcomes [116]

^fThe quality of evidence from the case-control study was downgraded from “low” to “very low” because of a serious risk of bias that diminished the level of confidence in the observed effects

^gIncludes 16 cross-sectional studies [90, 94, 100, 104, 107–111, 114, 115, 117, 118, 121, 123, 124]. Zimmerman et al. [117] and Ferguson and Donnellan [124] reported data from the same sample. Results are presented separately and participants are counted only once

^hSerious risk of bias. Potentially inappropriate sampling technique resulted in a sample with higher income and education than the overall population from which it was recruited [117, 124]; questionable validity and reliability of the exposure measure [90, 106–109, 111, 115, 117, 121–124]; questionable validity of exposure measure [94]; validation study showed overestimation of TV time exposure measure [110]; questionable validity and/or reliability of the outcome measure [109, 110]; unknown amount [109, 117] or between 28% and 60% [121, 124] of unexplained missing data and pattern of nonresponse indicates reason for missing data may have been related to the outcome of interest; incomplete reporting of exposure [109] and outcome [90, 110]; longitudinal relationships were reportedly collected but not reported in the results [115]; the methods section of one study indicated that bivariate analysis would be performed, but included variables and the results of the analysis were not reported [121]

ⁱThe quality of evidence from longitudinal studies was downgraded from “low” to “very low” because of a serious risk of bias that diminished the level of confidence in the observed effects