The primary goal of the present study was to examine whether autistic traits in a general population would be associated with within- and/or between-system FC among the three attentional systems, i.e., the Alerting, Orienting, and EC systems. There were significant associations between AQ scores and between-system FC values but not between AQ scores and within-system FC values; the AQ scores had positive relationships with FC between the regions responsible for Alerting and Orienting and with FC between the regions responsible for Orienting and EC, but a negative relationship with FC between the regions responsible for Alerting and EC.
In the present study, a greater degree of autistic tendencies was related to stronger FC between the thalamus, which is the center of Alerting, and the fusiform gyrus responsible for Orienting (i.e., the FFA). Previous studies have consistently reported that orienting is impaired in individuals with ASD (for a review, see Keehn et al. 2013a) and that this function is weaker in individuals with a high level of autistic traits from a general population (Nummenmaa et al. 2011; Zhao et al. 2015). Alerting, or the function of achieving and maintaining a state of sensitivity to incoming information, promotes orienting (Callejas et al. 2005; Fuentes and Campoy, 2008; Spagna et al. 2014). Thus, the present findings that there was stronger FC between Alerting and Orienting in individuals with higher levels of autistic traits might suggest weaker function in Orienting might be compensated for by Alerting.
Based on evidence that the FFA is an important factor in ASD and non-clinical individuals with higher levels of autistic traits, the notion that this region might be involved in weaker function in Orienting is reasonable. The FFA is known to be specialized for face perception (e.g., Kanwisher et al. 1997). Individuals with ASD exhibit impairment in face perception (for a review, see Weigelt et al. 2012) and the functionally and structurally atypical FFA (Schultz et al. 2000; Pierce et al. 2001; Waiter et al. 2004; Dziobek et al. 2010). Additionally, a positron emission tomography (PET) scan study revealed that cholinergic deficits in the fusiform gyrus, which are related to impairments in social interaction, are evident in subjects with ASD (Suzuki et al. 2011). Because the Orienting system, but not the other attention systems, is cholinergic (Petersen and Posner. 2012), these findings suggest that orienting deficits in individuals with ASD might be partly due to differences in the FFA and that higher levels of autistic traits might be associated with more deficient Orienting. Because individuals with subclinical but high levels of autistic traits also exhibit diminished activation in the fusiform gyrus during a face recognition task (Dalton et al. 2007), people with higher levels of autistic traits might have a weaker orienting function that is related to alterations in the FFA.
The weaker orienting function, associated with autistic traits, could also have impacted FC between Orienting and EC. The present results revealed that a greater degree of autistic tendencies was related to stronger connectivity between the precentral gyrus (Orienting) and the inferior frontal gyrus (IFG; EC). The function of the precentral gyrus in Orienting is thought to be similar to that of the frontal eye field (Fan et al. 2005), which implements eye movements such as saccades. Saccades are the phenomenon of initiating rapid ballistic shifts in eye gaze that are needed for attentional shifts during the orienting process (Awh et al. 2006). Several studies have reported that saccades are impaired in individuals with ASD (for a review, see Mosconi and Sweeny 2015). Moreover, because saccades are also impaired in unaffected first-degree relatives of individuals with ASD (Mosconi et al. 2010), those with a greater degree of autistic traits might have weaker saccade function. On the other hand, Kane et al. (2001) found that people with higher levels of executive function perform better on a saccade task than those with lower levels of executive function, which suggests that EC improves the function of saccades. Taken together with functions related to the IFG, which involves the selection of information to adjust sensory inputs (Barbas et al. 2011; Yoshimura et al. 2017), the present results suggest that weaker saccade function in individuals with higher levels of autistic traits might be compensated for by EC.
In the present study, a greater degree of autistic tendencies was also related to weaker FC between the cerebellar vermis (Alerting) and the two EC regions, i.e., the fusiform gyrus (as mentioned in the Methods section, this region is regarded as a non-face area) and the IFG. This finding may indicate that there is a delayed development of the Alerting-EC relationship in individuals with a greater degree of autistic tendencies. In general, relative to Alerting and Orienting, EC follows a protracted development period that can persist through adolescence (for reviews, see Diamond 2002; Best and Miller, 2010). In such developmental course, while functional interactions among the attentional systems except the Alerting and EC interaction are characterized in childhood and persist into adulthood, the direction of the Alerting-EC interaction is generally reversed by 12 years of age such that the positive relationship in childhood changes into a negative one in adolescence (Pozuelos et al. 2014; Mullane et al. 2016). On the other hand, a behavioral study examining the interactions among the attention systems found that there is a positive association between Alerting and EC in adolescents with ASD (Keehn et al. 2010) whereas typically-developing adolescents have a negative association (Fossella et al. 2002) or lack an association (Keehn et al. 2010). Considering that adults with ASD do not exhibit an Alerting-EC association (Fan et al. 2012) and that EC skills mature at a slower pace in ASD individuals than typically-developing people (for a review, see Demetriou 2018), these findings reported by previous behavioral studies suggest that the functional reversal of the Alerting-EC relationship might be delayed in ASD and that the development of this relationship in adults with ASD might be in a stage equivalent to that observed in typically-developing adolescents. Moreover, taken together with the notion that age-related changes in the interactions of attentional behaviors are associated with changes in FC (Rueda et al. 2015) and findings that age is an important factor to consider when assessing FC alterations in ASD (Hull et al. 2017; Nomi and Uddin 2015), it is likely that individuals with ASD show delayed or different patterns of developmental changes in FC between Alerting and EC. The present results showing that higher levels of autistic tendencies were related to weaker FC between Alerting and EC may indicate that this type of atypical development also occurs in non-clinical individuals with high levels of autistic traits.
The present findings also have clinical implications regarding the mental health of non-clinical individuals with higher levels of autistic traits. A considerable number of studies have documented higher rates of psychiatric problems, such as anxiety and depression, in this non-clinical population (e.g., Piven and Palmer 1999; Micali et al. 2004; Wilcox et al. 2003); these issues could be derived, at least in part, from orienting dysfunction. In other words, these disorders could be due to attention biases, such as reduced attention to positive information, excessive attention to negative information, and local processing biases; such kinds of biases could induce a distressed mood (Shechner et al. 2012; Armstrong and Olatunji 2012; Bradley et al. 2000; Gasper and Clore 2002). Because this type of orienting dysfunction has been reported in non-clinical individuals with higher levels of autistic traits (Losh et al. 2009; Sato et al. 2017), it may be a cause of the elevated rates of other psychiatric conditions in this population. Moreover, if Alerting and EC compensate for Orienting in this population, as suggested by the present findings, then training Alerting and EC functions might result in maintenance of mental health through functional improvements in orienting. The function of attention networks, especially that of EC, can be improved by changes in brain states induced by exercise (Hillman et al. 2014; Erickson et al. 2015) or meditation (Tang et al. 2014). Thus, investigations of the relationship between training attentional functions through these activities and mental health in this population would likely to provide interesting results.
The present study has several limitations that should be noted. First, the present findings are limited by the age of the participants. As described above, it is possible that the directions of the interactions among the attention systems might be partly reversed with increasing age. Future research using elderly populations will clarify whether this functional reverse might happen and/or when it does. Second, the relationships between actual attentional functions and the between-system FC of attention networks were not assessed in the present study. Although resting-state brain networks resemble task-evoked networks (Tavor et al. 2016), the relationships among the three attention systems (i.e., between-system FC) might change during active attention. Therefore, task-dependent FC should be studied using a combination of imaging and behavioral data to reveal possible changes in the relationships among active attention systems and in relation to the presence of compensatory mechanisms. Third, future investigations should consider within-system FC. In the present study, autistic traits were not associated with within-system FC whereas stronger and weaker FC within Orienting have been reported in individuals with ASD (Keehn et al. 2013b; Farrant and Uddin, 2016). It is possible that the present results were influenced by the non-clinical nature of the participants or the relatively small sample size. Thus, studies with larger sample sizes should be conducted to determine whether autistic traits in the general population are associated with within-system FC.