The physical environment encompasses all aspects of a child’s physical world and may be defined as objective characteristics of the physical context in which children spend their time (e.g., home, neighbourhood, school). The influence of children’s physical exposures has been summarised differentially by various models, theorems, and theorists over the previous century. Notably, these include the physical environmental elements of children’s exposome (a term introduced by Wild ^[1][2][1,2] regarding the non-genetic influences on outcomes across the lifespan) and Bronfenbrenner’s bioecological model ^[3][4][5][3,4,5], proposing that children develop within an environmental milieu of five interconnected systems, spanning aspects from urban design (e.g., presence and structure of sidewalks), traffic density, and design of venues for physical activity (e.g., playgrounds, parks, and school yards), to biologically active chemicals, radiation, the internal chemical environment, and psychosocial aspects [6]. The difficulty with these conceptualisations of child development is that they include both physical environmental and (psycho)social influences. As exemplified in a review of the influence of interior hospital environmental interior conditions, Harris [7] segmented the environment into distinct physical exposure categories: ambient, architectural, and interior design.

An operationalised definition of “physical environment” was incorporated to identify eligible environmental exposures. This classification was derived from a coalescence of Harris’s [7], Bronfenbrenner’s ^[3][4][5][3,4,5] and Wild’s ^[1][2][1,2] theorems. This resulted in the inclusion of a diverse array of domains, from ambient exposures (sunlight, sound, meteorology), interior design elements (colour, lighting), architectural features (space/spatial crowding), and biological active agents (i.e., air particulate pollutants), to physical aspects of children’s microsystem (i.e., home, school, and neighbourhood characteristics).

A variety of theories have attempted to explain the mechanisms via which environmental domains influence physical and mental health. Although none of these mechanistic models have been fully proven, there are suggestions that positive effects may be the end product of pathways that link several elements, such as mitigation (reduction in air pollution or traffic noise), restoration (stress reduction and attention restoration in alignment with what the Attention Restoration Theory posits) and instoration, whereby attributes of the physical environment, such as greenness in particular, may promote physical activity and social capital and cohesion ^[8][9][10][8,9,10].

Previous theories have primarily focused on the stress-reducing effects of greenspace, either via a protective influence from harmful environmental stimuli (noise and air pollution) ^[11][12][13][11,12,13], or via the restoration of attentional resources ^[14][15][14,15]. Recently, it has been posited that greenspaces may provide more direct physiological benefits via increased exposure to phytoncides (plant-derived antimicrobial volatile organic compounds) [16]. Whilst preliminary research into the effects of phytoncide exposure is positive, it is currently inconclusive and additional studies are required [17]. Neuroimaging studies are also shedding insight into potential mechanisms for greenspace exposures potential mechanisms, with one study [18] showing that it can beneficially deactivate the prefrontal cortex in regions linked to depression and rumination.

The literature also indicates that aspects such as ambient air particulate matter exposure may negatively impact development via neuroinflammatory pathways ^{[19][20][21][22][23][24]}[19,20,21,22,23,24]. Noise pollution may also have detrimental effects via contributions to subjective annoyance and irritation; whilst not necessarily directly causing aggression, noise exposure in those with low threshold for expressing anger may increase its severity ^[25][26][25,26] via draining of attentional and cognitive resources and subsequently leading to increased self-regulatory difficulties [27]. Social-behavioural mechanisms may explain the relationship between behaviour and fluctuations in meteorological effects (such as temperature), e.g., the routine activity theory that proposes that warmer temperatures facilitate more frequent social interaction, increasing opportunity for aggression [28] or that heat increases hostility and physiological arousal and consequently to aggressive behaviour [29]. Theories have posited that high spatial density triggers perceptions of crowding and a subsequent physiological stress arousal response ^[30][31][32][30,31,32]. Why proximity elicits these responses is still unclear and has been linked to competition for resources and invasion of personal space [33]. Baird et al. [34] reported a beneficial association between household crowding and reduced conduct problems in children with intellectual disabilities. The authors propose several theories about these potentially counterintuitive findings, suggesting that increased availability of and proximity to family members, in intergenerational households, and parental habituation to problematic conduct behaviours are all potential mechanisms underpinning this finding. Using a sensory room unaccompanied may be associated with a sense of autonomy in children and young people which in turn reduces distress [35]. Other pathways may contribute to the impact that music listening has on a broad range of psychological and physiological benefits ^{[36][37][38][39][40][41]}[36,37,38,39,40,41].

As discussed, social-behavioural mechanisms may explain the relationship between aggression and climate effects, for example the routine activity theory proposes that warmer temperatures facilitate more frequent social interaction, increasing opportunity for aggression [28]. Alternatively, the general aggression model (GAM) is more grounded in a physiological aetiology of aggression, suggesting that heat increases hostility and physiological arousal and consequently aggressive behaviours [29].

From the evidence presented so far, it appears that both physical and social environments, in addition to genetic and epigenetic influences, shape the developmental trajectories of children ^{[42][43][44][45][46]}[42,43,44,45,46]. However, in the main, published research is focused on typically developing rather than neurodiverse children [47]. Previous work has evidenced disproportionate influence of children’s early environmental milieu in shaping a range of socio-emotional and cognitive developmental outcomes. Specifically, learning disabled children are more likely to be affected by social adversity, poor housing, and poverty [48]. These children are also exponentially more likely to be exposed to negative environmental exposures such as air pollution [49]. To address failings in supporting these children and their families, an important element is to reduce socioeconomic inequality and improve residential conditions [50]. Furthermore, children with complex neurodisabilities have increased barriers to accessing potentially therapeutic aspects of both the physical and social early environments [51]. Disabled child access to urban greenspaces, for example, is not only infrequent in comparison to their typically developing peers [52], but when significant resources are employed to facilitate access for neurodisabled children, the high-risk nature of visiting these spaces requires rigid structure, impacting on the quality of nature experiences when they do occur [53]. This is one example of the health inequities and disparities experienced by neurodiverse children in comparison to their peers, exemplifying the need for additional research in these domains.

Externalising disorders are characterised by display of a range of behaviours which are associated with poor impulse-control, and include rule breaking, impulsivity, and inattention; in addition, a core component of these conditions is the presence of heightened aggression.

Specific child and adolescent externalising disorders include conduct disorder (CD), oppositional defiant disorder (ODD), and attention- deficit-hyperactivity disorder (ADHD). Of particular concern is this repeated presence of aggressive behaviour in these disorders as it is often associated with referral to services and application of a range of restrictive practices, most commonly antipsychotic medications but also inpatient admissions.

Aggressive behaviours and general behavioural problems such as destructive behaviours have an overall negative influence on carers due to stress and negative interactions between carers and the person they care for, likely resulting in a deterioration of the quality of care [54]. Moreover, behavioural problems are associated with increased service costs because of the impact of behaviours on staff and need for high support levels [55]. Aggressive episodes also provoke concerns about threat to personal safety as well as cause panic and upset ^[56][57][56,57].

These behaviours in both typically developing and neurodiverse children compound societal and educational limitations ^{[58][59][60][61]}[58,59,60,61]. They reduce life satisfaction via degradation of social and familial relationships [62], increase economic costs [63], require higher use of physical restraints [64] and restrictive environmental placements ^[65][66][65,66], limit access to support services [67], impair caregiver functioning ^[68][69][68,69], reduce educational opportunities due to teacher burnout [70] and encourage use of restrictive practices including psychotropic medication use ^[71][72][71,72].

2. Greenspace

Eleven longitudinal and seven cross-sectional studies (~46,684 participants) examined associations between greenspace exposure and childhood aggression ^{[34][73][74][75][76][77][78][79][80][81][82][83][84][85][86][87][88][89]}[34,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122]. Five studies were carried out in the UK, four in the USA, two in Belgium, with the remaining in Australia, Korea, Lithuania, Germany, Spain and China. All greenspace studies were classified as low RoB. Inconsistent evidence for harms or benefits was reported across eight studies ^{[73][74][75][76][79][80][84][88]}[106,107,108,109,112,113,117,121] that examined associations between satellite derived neighbourhood greenspace (NDVI) and parental-reported child aggression related outcomes. Two studies ^[76][90][101,109] examining the association between parental-reported child aggression and conduct problems and percentage of land designated as natural land, reported high-quality evidence. Proximity of the child’s residence to greenspace was inconsistently associated with parent reported conduct problems across three studies ^[73][75][81][106,108,114]. Very low-quality evidence ^[78][83][87][111,116,120] reported no relationship between percentage of neighbourhood greenspace and both child and parent-reported conduct problems. Moderate-quality evidence from three studies ^[34][78][87][34,111,120] reported inconsistent beneficial effects of access to private garden space on parent-reported conduct problems.

3. Environmental Sound and Noise

Three longitudinal and eight cross-sectional studies (n = 23,665) assessed the association between environmental noise pollution including road traffic, construction noise, aircraft noise and aggression outcomes ^{[91][92][93][94][95][96][97][98][99][100][101]}[123,124,125,126,127,128,129,130,131,132,133]. These studies were primarily conducted in the UK, Spain, Germany, and the Netherlands, and one study in China. Three of these studies ^{[92][99][100]}[124,131,132] used data from the multi-national RANCH study examining the influence of high and low road and aircraft noise on the behaviour of pupils who attended schools that were close to main roads or under flypaths. Two studies were judged to be of unclear RoB ^[94][98][126,130], with the majority being rated as low RoB. A very low-quality evidence for harmful association ^[98][130] between residential aircraft noise exposure and increased child annoyance was reported. Similarly, low- and very low-quality evidence was found for associations between increased residential noise ^[94][126], predicted air and road traffic noise ^[99][131], and heightened self-reported child annoyance. Schools located in areas of high aircraft noise were associated with increased child-reported annoyance ^[95][96][97][127,128,129], but inconsistently correlated with parent-reported child conduct problems (^[95][97][127,129], both very low quality). Two studies ^[92][100][124,132] examining the role of residential aircraft noise on the parent-administered conduct problems subscale of the SDQ reported no association (low quality). Five studies ^{[91][92][93][101][102]}[123,124,125,133,185] reported very low quality inconsistent evidence for estimated noise exposure effects on parent-reported child aggression. Two studies (longitudinal and within-group repeated measures) from the USA (n = 658) assessed the association of music on childhood aggression ^[103][104][183,184]. Both studies were rated as high RoB. Aggressive or sexual music content was associated with increased self-reported aggressive behaviour in adolescents (^[103][183], very low quality). Low-quality evidence reported no association between alternating periods of instrumental music and observer rated aggressive behaviours ^[104][184]).

4. Air Pollution

Eight longitudinal and six cross-sectional studies from Lithuania, China, Korea, Iran, Canada, USE, and the UK (n = 45,607) explored the influence of air particulate matter on aggressive behavioural outcomes in typically developing children and young people ^{[34][74][83][91][105][106][107][108][109][110][111][112][113][114][115]}[34,107,116,123,135,136,138,141,142,144,145,146,147,148,149]. One study was rated as high RoB ^[112][146], one as unclear RoB ^[106][136], and the remaining as low RoB. Five studies ^{[105][106][112][113]}[135,136,146,147] provided either low- or very low-quality evidence supporting the harmful influences of tobacco smoke exposure across various aggressive behavioural questionnaires. Very Low-quality evidence for a harmful association ^[100][132] between active or passive tobacco exposure and child self-reported anger and aggressive behaviour was found. Three studies ^{[83][108][115]}[116,141,149] examined the relationship between Nitrogen Dioxide (NO2) exposure and child self-reported conduct problems symptoms and reported inconsistent evidence for a harmful association (Very Low quality). No effect of Elemental Carbon Attributed to Traffic (ECAT) on parent-reported externalising behaviours (BASC-2) was found (^[111][145], Moderate quality). In addition, there was inconsistent evidence for an association between exposure to particulate matter less than 2.5 microns (PM2.5), and child self-reported conduct problems ^[108][115][141,149]. No effect was found in a study that examined the influence of PM2.5 on parent reported conduct problem scores ^[74][107]. Another study by Loftus et al., 2020 ^[110][144] explored the influence of exposure to particulate matter less than 10 microns (PM10) on parent reported child aggressive behaviours but it did not show a significant association (Very Low quality). Ambient air lead exposure (PbA)) was associated with high parent-reported aggressive behaviour (^[114][148], Very Low quality).

5. Meteorological Exposure

Five longitudinal and two cross-sectional studies (approximately = 6314) from Chile, Canada, the Netherlands, USA, and Italy, assessed associations between meteorological variables and child aggression outcomes ^{[116][117][118][119][120][121][122][123][124]}[151,152,153,154,155,156,157,158,159]. One study was rated as high RoB ^[123][158] with the remaining studies rated as low RoB. The study by Muñoz-Reyes et al. ^[123][158] contrasted the frequency of observed aggressive behaviours during the warm season (summer/spring) with the frequency of such behaviours during the cold season (autumn/winter), reporting Very Low-quality evidence for harmful effect of warm seasonality. Low-quality evidence associated increased humidity with harmful increases in teacher-reported child aggressive behaviours ^[116][121][151,156]. Studies examining the effects of sunlight exposure on teacher reported ^[116][121][151,156] and child self-assessment ^[117][152] behavioural outcomes reported inconsistent or no evidence, respectively (very low quality). Low- and Moderate-quality evidence for the harmful influence of increased temperature on teacher and parent-reported child aggression symptoms was reported in three studies ^{[116][121][124]}[151,156,159]. However, researchers found one Very Low-quality study that provided evidence for beneficial effect of temperature on children’s self-reported anger ^[117][152]. Aggression during summer recess was lower compared with aggression during the school year (^[119][154]; Low quality). No association between hours of precipitation per day and children’s self-reported anger was found (^[117][152], Very Low-quality evidence). Finally, a study carried out by Lochman et al. ^[118][153] examined longitudinal associations between tornado exposure and externalising symptoms, and reported a harmful associations of Moderate quality.

6. Spatial Density and Interior Design

Four observational and two longitudinal studies ^{[34][125][126][127][128][129]}[34,160,161,162,163,165] (n = 8568) from the USA and the UK examined spatial density and architectural design in relation to childhood aggression. RoB was judged as high in all studies except one rated as unclear ^[128][163] and one rated as low [34]. A study ^[129][165] reported a beneficial effect of increased playroom openness, but no effect of space per child or room group size on observed aggressive behaviours. Low-quality evidence assessing the association between high density (in comparison to low density) child playrooms and frequency of aggressive behaviours reported inconsistent results ^{[125][126][127]}[160,161,162]. Moderate-quality evidence examining the effect of overcrowding in the home ^[34][128][34,163] reported inconsistent associations with parent-reported conduct problems but was associated with reduced teacher-reported externalising behaviours. Three studies, two quasi-experimental and one longitudinal (n = 8257) conducted in Iran, the UK, and the USA examined the associations between interior design features and childhood aggression ^{[34][130][131]}[34,179,181]. Low-quality evidence of association (^[131][181] unclear RoB) between red painted classroom walls and increased self-reported aggression was found. In-patient psychiatric ward sensory room modifications were correlated with beneficial reductions in observer rated aggressive behaviour (^[130][179] unclear RoB Moderate quality). Additionally, presence of damp in the house was associated with elevated trajectories of conduct problems in children ([34], Moderate quality).

7. Urbanicity and Rurality

Three longitudinal and three cross-sectional studies (n = 17,630) from the USA, the Netherlands, and Thailand explored the influence of urbanicity and rurality of residence on children’s aggressive behavioural outcomes ^{[34][132][133][134][135][136]}[34,164,168,170,172,176]. One study was rated as high RoB ^[133][168] with the remaining assessed as low or moderate RoB. One study ^[136][176] reported inconsistent associations between the location of the participants and scores across three self-reported aggression outcomes (Very Low quality). Moderate-quality inconsistent evidence ^[34][135][34,172] was reported for the effect of urban residence on child conduct problems and aggressive behaviour in parent-reported questionnaires, whilst evidence for a lack of association was found for teacher-completed aggression outcomes ^[135][172]. Another study ^[134][170] examined the effects of urban or rural settings on aggressive behaviours in schoolchildren attending schools from either setting. It reported no association of setting with parent-reported behaviours, but a harmful effect of urban school location on teacher-assessed behaviours (both Very Low quality). Very Low-quality evidence reported no association between children recruited from rural or urban Head Start centres and teacher-reported anger ratings (^[133][168]: AML Behaviour Rating Scale). Neighbourhood urbanicity (mean number of addresses within a 1 km radius of participant’s residence) was associated with increased teacher-reported child problem behaviours (^[137][166] Low quality).