Visual attention is one of the most robust early indicators of autism spectrum disorder (ASD), with higher rates of atypical features identified prior to age 12 months in infants later diagnosed with ASD [
1,
2]. The majority of studies of early attention in ASD have been conducted with “high familial risk” (HFR) infant siblings of children with ASD who exhibit nearly 20 times the risk of ASD diagnoses than the general population [
3]. Although a number of abnormal attention processes in HFR infants have been identified as potential “red flags” for ASD, the emergence and mechanisms of these behaviors remain unclear. Integrating biological measures of attention into prospective HFR infant studies may inform more nuanced developmental trajectories, clarifying the onset and course of atypical attention in HFR infants and potentially advancing early screening or treatment protocols. In infants, decelerations in heart rate offer one such biological measure that can be used to measure the presence and quality of “sustained attention,” the process whereby an infant exerts cognitive resources to process a stimulus [
4‐
8]. Indeed, because heart rate deceleration has been shown to provide a more sensitive metric of sustained attention than behavioral looking alone [
9], this method has been successfully applied to measure attention-related outcomes in a number of recent clinical infant nutrition studies [
10,
11]. As such, it is possible that heart rate-defined sustained attention (HRDSA) may be similarly applied as a biological measure to more sensitively index attention in HFR infants. However, given self-regulatory deficits commonly reported in ASD [
12,
13], it is unclear whether HRDSA will present similarly in HFR infants and previously studied low-risk populations. Thus, the present study aims to preliminarily characterize the nature and clinical correlates of HRDSA in HFR infants, informing whether HRDSA may operate as a promising biological measure of sustained attention and relevant clinical features—such as emergent ASD features—in this population.
Heart rate-defined sustained attention
It is well-established that infants’ physical look durations do not cleanly map onto attention, as infants will often continue looking when they are no longer cognitively engaged in stimulus processing [
14]. Thus for over five decades, heart rate-defined sustained attention has been integrated into developmental research as a well-validated, low-cost, and non-invasive biological measure of attentional engagement [
15]. When the brain’s arousal system is activated, cardioinhibitory centers initiate parasympathetic processes to slow heart rate, producing heart rate decelerations that quantify qualities of stimulus engagement that cannot be measured using overt looking patterns alone [
4‐
7]. Richards and colleagues have quantified three primary attention phases that occur when infants are physically looking toward a stimulus: orienting, sustained attention, and attention termination [
4,
7,
16‐
18]. Heart rate-defined sustained attention (HRDSA) is indexed by maintenance of decelerated heart rate, reflecting the exertion of additional cognitive resources to process a stimulus. Indeed, infants in periods of HRDSA are less distractible across a variety of tasks, including computerized orienting paradigms [
5,
19,
20] and toy play activities [
21,
22]. HRDSA is also associated with enhanced neural responses and more accurate performance on behavioral learning tasks [
19,
23‐
25], supporting HRDSA as a valid measure of attention in typically developing infants.
HRDSA in special populations
Few studies have examined heart activity in HFR infants, and it is possible that HRDSA may present differently in HFR given the well-established autonomic abnormalities associated with ASD (see [
12,
13] for review). A number of studies of older individuals with ASD have generally suggested faster overall heart rate in ASD [
26‐
28] and difficulties modulating arousal when task demands change [
29‐
31], although inconsistent findings are common [
13]. Some evidence suggests that inconsistencies may reflect the broad, heterogeneous nature of ASD symptoms, as studies specifically focused on social communication deficits within ASD—as opposed to broader ASD status—have generally reported significant associations between ASD and physiological arousal [
13]. In addition, it is possible that developmental factors complicate interpretation across studies. In a recent study, HFR infants exhibited slower heart rate (HR) at 3, 6, 9, and 12 months of age relative to low rate (LR) controls, conflicting with findings of hyperarousal in older individuals with ASD [
13]. Notably, this pattern of slower HR in infants at high risk for ASD parallels previous findings in children with fragile X syndrome (FXS), a single-gene disorder highly associated with ASD. Specifically, infants with FXS who later exhibit high ASD features display slower HR in infancy (< 12 months) yet faster HR in the toddler and preschool period (12–40 months), relative to infants with FXS and low ASD features [
32]. Given the complex, heterogeneous symptom profiles and developmental factors associated with ASD, further work is needed to characterize associations between heart activity and symptom profiles, particularly in early childhood.
Similarly, few studies have examined heart rate decelerations as a specific index of attention in ASD, and existing studies have used variable methods that preclude generalizable conclusions about the nature of HRDSA in ASD. Corona and colleagues examined behavioral responses and heart activity of 3- to 5-year-old children with ASD who viewed an examiner pretending to hurt herself, then displaying either an intense distress response or neutral affect [
33]. Behaviorally, children with ASD and comparison children with intellectual disability both looked more toward the examiner displaying distress, although the ASD group exhibited atypical qualities of social attention, such as fewer looks toward the examiner’s face. However, only the control group exhibited decelerations in heart rate when examining the examiner in distress, potentially indicating blunted HRDSA toward the emotional stimulus in the ASD group. In contrast, Louwerse and colleagues identified typical heart rate decelerations in adolescents with ASD relative to non-ASD controls during a picture viewing task, with both groups exhibiting larger heart rate decelerations to unpleasant versus neutral stimuli [
34]. Notably, both studies focused on heart rate decelerations as a metric of sustained attention in response to emotion-focused stimuli, rather than examining how HRDSA may inform general sustained attention in ASD. Thus, it remains unclear how HRDSA may inform general sustained attention and stimulus processing in ASD.
More recent work has begun expanding studies of HRDSA to infants at risk for ASD. In a longitudinal pilot study of 9- to 18-month infants with FXS (
n = 13) and 12-month controls (
n = 10), infants with FXS generally exhibited shallower heart rate decelerations and lower heart rate variability during a visual attention task [
22]. However, despite infants with FXS displaying shallower decelerations at the group level, higher ASD symptoms were marginally associated with deeper heart rate decelerations and greater behavioral looking within the FXS group. Thus, despite an overall profile of physiological dysregulation in FXS, deeper heart rate decelerations converged with enhanced visual attention, similar to patterns observed in typically developing infant samples (e.g., [
5,
19]). More recently, atypical patterns of heart rate decelerations were identified in HFR infants relative to low familial risk (LFR) controls from 3 to 12 months of age in response to speech stimuli [
35]. Specifically, HFR infants exhibited atypically shallow heart rate decelerations across development, increasingly deviating from LR controls between 3 and 12 months of age, suggesting reduced social orienting during time. However, heart rate decelerations were not examined in relation to ASD outcomes; thus, it remains unclear whether atypically shallow decelerations observed in HFR infants are driven by the subset of HFR infants with later ASD. Together, these preliminary findings suggest that HRDSA may provide novel information about attention in infants at risk for ASD, although differences across stimuli, risk group, and availability of ASD outcome data complicate interpretations. For example, when HRDSA is used to measure social orienting (e.g., [
35]), which is reduced in ASD, ASD risk may be associated with shallower HRDSA. In contrast, when HRDSA is used to measure responses to non-social visual stimuli (e.g., [
22]), which are often reported to be atypically enhanced in ASD, ASD risk may be associated with increased HRDSA. Alternately, it is possible that differences in HRDSA patterns may be driven by biological or genetic factors, with FXS and HFR status mapping on to different patterns of physiological features. Given the limited work in this area, additional studies are needed to characterize the profiles of HRDSA in special populations, with particular attention toward intersecting developmental, clinical, and biological factors.
HRDSA and visual attention in HFR infants
Although studies of HFR infants have critically advanced understanding of the emergence and early symptoms of ASD in infants, a major challenge to this work is the vast heterogeneity in individual profiles and features that present in infants later diagnosed with ASD. Biological measures such as HRDSA may improve HFR infant research by supporting more precise measurement of individual differences, reducing error and increasing power in HFR infant attention studies. Indeed, although abnormal cognitive and social attentional features have been observed in HFR infants and have been the topic of several reviews [
36,
37], attention in HFR infants remains a complex topic, in part due to the diverse conceptualization and measurement of attention in the extant HFR infant literature. It is possible that physiologically derived indices of sustained attention could advance HFR infant attention research by providing information about the quality of attentional responses, particularly when looking behaviors are a primary variable of interest.
Relevant to the present study, a number of studies suggest abnormal attentional patterns in HFR infants across a variety of visual attention tasks, with particularly atypical profiles among infants with later ASD features. For example, HFR infants who later meet ASD diagnostic criteria exhibit longer saccadic latencies to disengage attention from competing stimuli [
2,
38] and enhanced visual search performance [
39]. In addition, studies focused on more socially salient tasks have demonstrated reduced attention-related behaviors in HFR infants, with later ASD features associated with reduced attention toward social scenes [
40], people within a social scene [
40], and a person’s eyes within a social scene [
41] by 6 months of age. Fewer studies have examined sustained attention specifically in HFR infants, with one recent study suggesting disrupted social sustained attention in HFR infants who later meet criteria with ASD (
n = 9) who exhibited shorter peak looks on social trials of a habituation task compared to HFR infants who did not meet ASD criteria (
n = 49), although group differences were no longer significant at 12 months [
42]. Limited work in older individuals is similarly complex, with some studies reporting reduced sustained attention in preschoolers with ASD [
43], and others suggesting intact sustained attention in adolescents with high-functioning autism [
44]. Reflecting this complexity, no single attentional profile established in either HFR infants in general or the subgroup of HFR infants who later meet criteria for ASD [
45].
Improving the quality and depth of attention measurement during visual attention tasks, such as through the integration of HRDSA, may help clarify the scope, emergence, and mechanisms of abnormal attention in this population. At the phenotypic level, examining HRDSA may further inform whether sustained attention is disrupted in HFR infants, particularly those with later ASD features, building on a single study that has explicitly addressed this topic to date [
42]. Methodologically, HRDSA could also be used to more sensitively capture the quality of attention engagement among HFR infants, as several recent studies have identified group differences in HRDSA despite similar behavioral sustained attention, such as in the context of habituation tasks [
9] and nutritional clinical trials [
11]. At the analytic level, HRDSA may also be used to reduce error in visual attention experiments by enabling researchers to isolate trials in which infants are physiologically engaged in an attention task. For example, typically developing infants exhibit more efficient stimulus processing and greater recognition memory when presented stimuli during HRDSA versus non-HRDSA phases [
46]. Thus, researchers may control for HRDSA when examining condition effects or alternatively design experiments to selectively administer trials once a period of HRDSA has been achieved [
21,
47]. Although other neuroimaging methods such as electroencephalogram (EEG) have similarly yielded more sensitive information about visual attention in HFR infants compared to behavioral methods [
48,
49], HRDSA is a particularly attractive method given the relative ease, affordability, and portability of collecting infant heart rate data relative to more invasive or expensive neuroimaging techniques.
The present study
Heart-rate defined sustained attention is a promising biological measure for measuring and tracking the development of attention in HFR infants given both the popularity of this method in typical developmental literature and its emergent promise in characterizing phenotypes in infants with FXS [
22]. To test whether HRDSA may similarly inform attentional development in HFR infants, the present study examined levels and age-related changes in HRDSA in a prospective cohort of HFR infants and low-risk controls during a passive looking task. Although the focus of the present study is on HRDSA specifically, we expected age-related differences in task performance across groups, with LFR controls exhibiting reduced behavioral attention over time [
50]. We also expected the task to solicit greater behavioral attention in HFR versus LFR infants, with increasingly distinct profiles between 6 and 12 months, based on previous findings that HFR infants with greater ASD features exhibit increasingly poorer disengagement from visual stimuli [
1,
2] and neural changes predictive of ASD [
51] between 6 and 12 months. In the context of these expected behavioral patterns, our goals were to characterize profiles of HRDSA in HFR infants relative to LFR infants, controlling for behavioral attention, as well as determine whether HRDSA corresponded with indices of emergent ASD features in the HFR group. This approach contributed to our overall aim of determining whether HRDSA may serve as a sensitive biological measure of sustained attention in HFR infants.
Given the exploratory nature of this study, we aimed to test two distinct hypotheses regarding the nature of HRDSA in HFR infants. The first hypothesis was that HFR infants would display greater HRDSA, indexed by greater proportion of time in HRDSA and deeper heart rate decelerations during HRDSA, and that these patterns would be most atypical in infants who later exhibited more severe ASD features. This hypothesis reflects a global assumption that HRDSA would relate to behavioral sustained attention similarly across HFR and LFR groups, with greater HRDSA paralleling greater behavioral attention, similar to previous studies of typically developing infants [
5,
19]. Indeed, older children with ASD have been shown to exhibit typical increases in HRDSA toward emotional stimuli [
34] and decreased HRDSA that parallels reduced quality of social attention [
33], suggesting HRDSA may converge with behavioral responses in this population. Similarly, greater heart rate decelerations have been preliminarily associated with greater ASD symptoms in infants with FXS, despite atypically shallow decelerations in FXS relative to controls [
22]. Thus, this first hypothesis posited that HRDSA in HFR infants would operate in a qualitatively similar manner to LFR infant studies, potentially advancing the sensitivity of sustained attention measurement in HFR infant studies.
Our alternative hypothesis was that HFR infants, particularly those with greater emergent ASD features, would display enhanced behavioral attention yet
shallower heart rate decelerations during HRDSA, similar to enhanced behavioral attention and shallower heart rate decelerations previously observed in fragile X syndrome [
22]. This hypothesis would suggest that in Corona and colleagues’ previous study, adolescents with ASD may have failed to exhibit typical heart rate decelerations across conditions due to a dysregulated sustained attention response, rather than a blunted emotional response to the examiner’s distress. This hypothesis is also consistent with previous studies in which individuals with ASD exhibit hyperarousal and difficulty modulating arousal across changing task conditions [
29‐
31], as well as evidence that increasingly faster heart rate across toddlerhood predicts ASD symptoms in FXS [
32]. In other words, it is possible that even during infancy, HFR status may be associated with dysregulated sustained attention responses that manifest in blunted HRDSA, potentially in concert with emergent ASD features in a subset of infants. If upheld, the implications of this hypothesis are that HRDSA may provide insight into the mechanisms sustaining atypical attention profiles in HFR infants, rather than solely advancing sustained attention measurement in this group.