Sensory Features as Predictors of Adaptive Behaviors: A Comparative Longitudinal Study of Children with Autism Spectrum Disorder and other Developmental Disabilities

Abstract

Background:

Children with autism spectrum disorder (ASD) and other developmental disabilities (DD) exhibit sensory features that differ from their typically developing peers. Prior cross-sectional research has demonstrated significant associations between elevated sensory features and lower adaptive behavior scores, yet there is limited prospective research examining longitudinal associations.

Aims:

To examine the longitudinal prediction of early sensory response patterns (i.e., hyperresponsiveness, hyporesponsiveness, and sensory interests, repetitions, and seeking behaviors) to later adaptive behavior outcomes in children with ASD and DD.

Methods and Procedures:

Children with ASD (n=51) and DD (n=30) were seen at two time points (Time 1: M(SD)=5.6(2.5) years; Time 2: M(SD)=9.0(2.2) years). We used a series of regression models with both observational and parent-report measures of sensory response patterns, and group interactions.

Outcomes and Results:

All three sensory response patterns significantly predicted aspects of adaptive behaviors, with some differences based on assessment format and diagnostic group. Across groups and sensory patterns, we found some evidence that elevated sensory features early in childhood predicted lower adaptive behavior skills later in childhood.

Conclusions and Implications:

Sensory features may interfere with the development of adaptive behaviors, suggesting a need for effective interventions addressing sensory features early in development.

1. Introduction

Children with autism spectrum disorder (ASD) and those with other developmental disabilities (DD) are frequently reported to present with unusual behavioral responses to sensory stimuli (i.e., sensory features) compared to typically developing children (Baranek, David, Poe, Stone, & Watson, 2006). Cross-sectional studies suggest that sensory features may be an important factor affecting daily life activities (Baker, Lane, Angley & Young, 2008; Dunn, 2007; Tomchek, Little & Dunn, 2015; Watson et al., 2011) and quality of life (Ashburner, Ziviani & Rodger, 2008; Cermak, Curtin & Bandini, 2010) for children with ASD and DD, as well as the well-being of their caregivers (Kirby, White & Baranek, 2015b). Therefore, many therapeutic and educational services aim to target these issues for better functional outcomes. However, there is little prospective research depicting the impact of early sensory features on later adaptive behaviors in these two populations. The present study focused on the longitudinal prediction of three sensory response patterns (i.e., hyperresponsiveness, hyporesponsiveness, and sensory interests, repetitions and seeking behaviors) early in life to adaptive behavior outcomes for children with ASD and DD. We also explored the extent to which diagnostic group may influence these longitudinal effects.

1.1. Sensory Features

Sensory features have been organized as multi-dimensional patterns of behavioral response to stimuli across modalities, including hyperresponsiveness (i.e., exaggerated, aversive or avoidant responses to stimuli), hyporesponsiveness (i.e., muted, delayed or absent responses to stimuli), and sensory interests, repetitions, and seeking behaviors (i.e., intense fascination or repeated engagement with specific sensory qualities of stimuli or sensory-based actions with the body). These behavioral patterns manifest very early in development (Baranek, 1999a; Green, Ben-Sasson, Soto & Carter, 2012), and persist throughout childhood (McCormick, Hepburn, Young & Rogers, 2016).

Sensory features are well-characterized in the literature for children with ASD and DD (Baranek et al., 2006; Ben-Sasson et al., 2009; Cheung & Siu, 2009; Kern et al., 2006). Cross-sectional studies indicate that many children with ASD experience both higher rates of sensory features (Baranek et al., 2006; Tomchek & Dunn, 2007; Wiggins, Robins, Bakeman & Adamson, 2009) and more complex responses than those with DD (Ben-Sasson et al., 2009) or typical development (TD; Ahn, Miller, Millberger & McIntosh, 2004; Ben-Sasson, Cermak, Orsmond & Tager-Flusberg, 2007; Dunn, Myles & Orr, 2002). Individuals with DD are also reported to have elevated levels of sensory features compared to the TD population (Cheung & Siu, 2009; Ermer & Dunn, 1998), thereby suggesting that developmental delay may be a contributing factor to the behavioral manifestation of sensory features across many disability groups.

1.2. Adaptive Behaviors

Adaptive behaviors comprise age-appropriate abilities for living and functioning in daily life including socialization, communication, daily living skills, mobility and community participation. Studies of children with ASD have found a pattern of relative deficits in socialization and communication compared to daily living skills (Carter et al., 1998, Freeman, Del’Homme, Guthrie & Zhang, 1999; Rodrigue, Morgan & Geffken, 1991; Volkmar et al., 1987), whereas the opposite trend has been reported for children with DD (Ditterline, Banner, Oakland & Becton, 2008; Dykens, Hodapp & Evans, 1994). Furthermore, deficits in adaptive behaviors for individuals with ASD may be disproportionately large relative to their cognitive abilities or verbal IQ (Klin et al., 2007), indicating particular importance in addressing these skills through various services.

The results of longitudinal studies tracking changes in adaptive behaviors in children with ASD and their comparison to children with DD are mixed. While some have shown adaptive behaviors to be stable from preschool to adolescence (e.g., Schatz & Hamdan-Allen, 1995), others have reported a decline in standard scores as children age (e.g., Fisch, Simensen & Schroer, 2002). Some studies indicate that children with ASD showed mild to moderate gains in their skills over time, but have a slower rate of growth compared to their peers with DD or TD (Anderson, Oti, Lord & Welch, 2009; Di Nuovo & Buono, 2007; Howlin, Mawhood & Rutter, 2000; Matson, Dempsey & Fodstad, 2009; McGovern & Sigman, 2005). Furthermore, children with ASD tend to show more gains in adaptive skills during early childhood compared to later years (Shattuck et al., 2007). Likewise, individuals with higher IQ scores appear to show more gains in adaptive skills than those with lower IQ scores (Baghdadli et al., 2012; Kanne et al., 2011). It is important to note that changing diagnostic criteria or varying measures of adaptive behavior may account for some of the heterogeneity in these findings.

1.3. Associations between Sensory Features and Adaptive Behaviors

Several cross-sectional studies have indicated negative associations between sensory features and adaptive behaviors. Rogers, Hepburn and Wehner (2003) found that sensory responsivity, as measured by the Short Sensory Profile (McIntosh, Miller, & Shyu, 1999), accounted for significant variance in overall adaptive behavior scores in a study of young children with ASD, after accounting for both developmental level and autism severity. Other studies have identified associations between individual sensory patterns and adaptive behavior scores. Among children with ASD and DD, Watson and colleagues (2011) found that higher hyporesponsiveness scores were associated with poorer social skills. However, they found that higher hyperresponsiveness tended to be associated with greater social skills. Other studies have consistently seen patterns of negative prediction between all three sensory patterns and adaptive behavior outcomes among children with ASD. Lane, Young, Baker and Angley (2010) found that a combined measure of hyporesponsiveness and sensation seeking negatively predicted communication in ASD. Liss, Saulnier, Fein, and Kinsbourne (2006) indicated that hyperresponsiveness and hyporesponsiveness negatively predicted social skills, and hyporesponsiveness also negatively predicted daily living and communication skills in children. Inclusion of a DD comparison group helps to distinguish whether associations are broadly relevant to children presenting with sensory features, or if the relationships differ depending on diagnosis. Only Watson et al. (2011) included a DD group for comparison.

Although the cross-sectional literature affirms that sensory response patterns may be differentially associated with specific aspects of adaptive behavior in children with ASD and DD, the nature of these associations over time remains unclear. Watson and colleagues (2011) stressed the importance of a developmental perspective, positing that intact sensory processing early in life is critical for development of adaptive behaviors. Specifically, detecting and orienting to salient sensory stimuli provides a mechanism by which infants and young children begin to engage with people and objects in the environment. Thus, dysregulated sensory experiences early in life could derail learning opportunities if stimuli are confusing, aversive, and/or insufficiently salient. As the expectations of daily life skills within the general population increase in number and complexity with age, deficits in individuals with ASD and DD may compound over time. Conceptually, therefore, atypical sensory features early in childhood may have an increasingly negative impact on daily functioning ability as children age (Watson et al., 2011). To our knowledge, only one study has used longitudinal methods to examine associations between sensory features and adaptive behaviors. McCormick and colleagues (2016) did not find significant associations between sensory features and adaptive behaviors over time among individuals with ASD after controlling for intellectual ability and symptom severity. However, because this study used an overall score from the Short Sensory Profile, and the overall adaptive behavior composite from the Vineland Adaptive Behavior Scales, it cannot inform the extent to which different sensory response patterns are predictive of different adaptive outcomes over time. Given the incongruence of the limited longitudinal findings with strong evidence from cross-sectional studies and theoretical arguments that sensory features may impact adaptive behavior outcomes, there is a need for more prospective empirical research, particularly using measures that can separate out the various patterns of sensory response as well as different domains of adaptive behavior.

The picture is complicated for longitudinal studies due to intervening variables over time. Individuals with ASD and DD often receive a variety of services throughout childhood that may target aspects of sensory features and/or adaptive skills (Howard, Sparkman, Cohen, Green & Stanislaw, 2005; Schaaf et al., 2014; Uyanik, Bumin & Kayihan, 2003). Services received often include therapy (i.e., occupational therapy, physical therapy, and speech and language therapy) and educational services. A few existing studies have addressed the effects of intervention on functional outcomes in individuals with ASD, although none have specifically targeted sensory related services or potential differences in outcomes for children with ASD compared to DD. Baghdadli and colleagues (2012) measured treatment dosage over 3 years as the average hours per week of intervention based on the therapeutic objective (e.g., communicative improvement, social interaction, motor ability) and reported that fewer hours of early intervention posed a risk for slower gains in communication skills. Another study noted that social skills improvement for children ages 2 to 13 was partially predicted by the amount of parent-mediated intervention that the children received (Anderson et al., 2009). Given the hypothesized positive influence of various interventions on development, longitudinal studies of behavioral outcomes in ASD and DD should take these types of services into account.

1.4. Study Purpose

A better preliminary understanding of the associations between early sensory features and later adaptive behaviors is necessary to build theoretical models for how specific mechanisms may influence downstream developmental consequences. Clarity regarding whether there are differential longitudinal associations between these constructs for children with ASD versus DD could also be useful in future design and tailoring of more effective intervention strategies. Therefore, this study aimed to examine the predictive effects of three sensory response patterns early in life (i.e., hyperresponsiveness [HYPER], hyporesponsiveness [HYPO], and sensory interests, repetitions, and seeking behaviors [SIRS]) for later adaptive behavior outcomes (i.e., daily living skills, socialization, communication, adaptive behavior composite) for children with ASD and DD. Based on conceptual arguments for the important role of early intact sensory processing for the ongoing development of adaptive behaviors (e.g., Watson et al., 2011) and previous cross-sectional studies with largely convergent findings, we hypothesized that higher levels of sensory features, particularly hyporesponsiveness, would predict deficits in later adaptive behaviors, most notably socialization and communication domains, in children in both groups.

2. Methods

This study used a longitudinal cohort design with two groups (i.e., children with ASD and DD) assessed at approximately age 5 and again at age 9. The study was approved by the institutional review board for the university and adhered to all recommended data security and informed consent/assent procedures.

2.1. Participants

Participants were enrolled in the Sensory Experiences Project, which consisted of two sequential five-year phases. Recruitment occurred through various sources, such as a state-wide research subject registry, schools, developmental evaluation clinics, and parent support groups. Families were compensated $20–$75 at each time point based on the number of assessments completed as part of the larger project. 50 children with ASD and 31 children with DD and their caregivers participated in the study across both time points [Time 1 (T1) age in years: ASD: 5.5(2.3) DD: 5.7(2.6); Time 2 (T2) age in years: ASD: 9.1(2.2) DD: 9.0(2.1)].

A full description of the sample characteristics is included in Table 1. Children in the ASD group were given a clinical diagnosis of ASD by an independent licensed psychologist or physician (e.g., developmental pediatrician, psychiatrist), which was confirmed in this study using the Autism Diagnostic Interview-Revised (ADI-R; Le Couteur, Lord, & Rutter, 2003) and Autism Diagnostic Observation Schedule (ADOS; Lord, Rutter, DiLavore, & Risi, 1999). Children included in the DD group comprised those with idiopathic developmental delays (n=11), known genetic syndromes (n=17), or delays related to prematurity (n=3). Children were excluded from the DD group if they were diagnosed with ASD or met cut-offs for autism on the ADOS or Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Renner, 1988). Exclusion criteria for both groups included: fragile X syndrome, tuberous sclerosis, uncontrolled seizure disorder, cerebral palsy, mental age <6 months, or failure on vision or hearing screenings.

Table 1.

Sample characteristics

Characteristic			ASD (n = 50)	DD (n = 31)
Age in years at T1—M (SD)			5.53 (2.28)	5.66 (2.62)
Age in years at T2—M (SD)			9.07(2.21)	9.01(2.11)
Years between time points—M (SD) IQ proxy—M (SD) Gender			3.54(1.51)	3.35(1.17)
			68.42(27.07)	59.43(18.74)
	Female		7	11
	Male		43	20
Race/ethnicity
	Asian race		1	0
	African-American race		3	3
	White race		42	26
	Other race(s)		4	2
	Hispanic ethnicity		9	1
Mother’s highest level of education
	High school graduate/GED		8	8
	Associate degree, technical training, or partial college		40	19
	Bachelor or advanced degree		2	4
Household yearly income
	< $60,000		12	15
	$60,000 – $99,999		23	9
	> $99,999		15	7

Notes. ASD, autism spectrum disorder. DD, other developmental disability. T1, first time point. T2, second time point.

2.2. Instruments

Trained research staff administered the assessments, generally over the course of two visits for each time point. Instruments used for the present analysis included two parent-report and two observational measures of sensory features (Time 1), a developmental assessment (Time 1), a measure of adaptive functioning (Time 2), and a structured parent interview detailing the history of all intervention services the child received between time points (Time 2).

2.2.1. Sensory measures.

Two well-validated (Ausderau et al., 2014; Kientz & Dunn, 1997; Little et al., 2011) caregiver questionnaires measuring the frequency of sensory features across modalities and contexts were administered at both time points: the Sensory Experiences Questionnaire Version 3.0 (SEQ; Baranek, 2009) consisting of 105 items, and the Sensory Profile (SP; Dunn, 1999) consisting of 125 items, both using a 5-point Likert scale. A small subset of children received an earlier version of the SEQ at the first time point. Both the SEQ and SP assess frequencies of a child’s unusual reactions to sensory stimuli across modalities.

Two observational assessments, the Sensory Processing Assessment (SPA; Baranek, 1999b) and Tactile Defensiveness and Discrimination Test-Revised (TDDT-R; Baranek, 1998) were also given at both time points. The SPA is a 20-minute semi-structured play-based assessment that provides opportunities to observe responses to sensory stimuli and interactions with sensory toys; validated in multiple studies of children ages 9 months through 12 years (e.g., Baranek et al., 2006; Baranek et al., 2013). The TDDT-R is a 20-minute assessment targeting measurement of tactile hyperresponsiveness and discrimination designed for children over 3 years of age, also using a play-based approach; it has been validated in numerous studies up to 14 years of age (e.g., Baranek, Foster & Berkson, 1997; Foss-Feig, Heacock & Cascio, 2012). Although focused on the tactile domain, this assessment includes toys that have multisensory properties (auditory, visual and tactile). Summary scores for sensory response patterns of HYPO, HYPER, and SIRS were calculated through an extensive and rigorous process that included (a) two researchers independently categorizing items as to the sensory response pattern each measured; (b) a consensus process to resolve any disagreements; and (c) confirmatory factor analysis for empirical verification. Item scores for all measures were transformed to a 1 to 5 scale and score valences were aligned (see Boyd et al., 2010 for more procedural details). For the current study, the summary scores for HYPO, HYPER, and SIRS from each measure were standardized to have the same mean and standard deviation, and then averaged for the two parent-report and two observed measures separately (for more information, see Watson et al. 2011).

2.2.2. Cognitive measures.

Research staff administered a standardized cognitive assessment, either the Mullen Scales of Early Learning (MSEL; Mullen, 1995), or the Stanford-Binet Intelligence Scales-5^th edition (Roid, 2003), to children in both the ASD and DD groups based on their age and developmental level,. A non-verbal mental age equivalent (NVMA) was calculated for each participant in order to avoid floor effects of standard scores, which was then transformed into an IQ proxy score. The IQ proxy score is a developmental quotient calculated as NVMA divided by chronological age, multiplied by 100 (Becker, 2003; Knobloch & Pasamanick, 1974), and was included as a covariate in the analyses.

2.2.3. Adaptive behavior measure.

The Vineland Adaptive Behavior Scales-Survey Edition (VABS; Sparrow, Balla & Cicchetti, 1984) is a standardized structured interview and was administered to a caregiver of each participant. Domain standard scores for Communication, Socialization, and Daily Living Skills, as well as an overall composite standard score (ABC), were used for analyses. We used the original version of this measure with all participants to maintain consistency with study data collected before the second or third editions were available.

2.2.4. Service usage.

A trained assessor conducted a detailed semi-structured parent interview at both time points to gather data on the amounts and types of services the children received. Data were collected on a wide range of services including medical, complementary and alternative, pharmacological, occupational therapy, physical therapy, speech therapy, and all educational interventions. Services were coded by specific type, start/stop dates, and duration, among other variables. The amount of services utilized between the two data collection time points that were relevant to the current analyses (i.e., therapy and educational services) were measured in number of hours per week and aggregated into two between-time point service usage variables: (a) total hours of traditional therapy services (i.e., occupational therapy, speech-language therapy, and physical therapy), and (b) total hours of specialized educational services (i.e., early intervention, applied behavior analysis, and school-based treatments, excluding occupational, speech-language, and physical therapy services).

2.3. Data Analysis

Data were double-entered and error-checked before analysis was conducted in Statistical Analysis Software, Version 9.3 (SAS Institute, 2011). The associations between the three sensory predictors (HYPO, HYPER, and SIRS) measured at T1 and adaptive behavior outcomes on the VABS (Daily Living Skills, Communication, Socialization, and the ABC) measured at T2 were assessed using a series of regression models. To examine group interaction effects, each model initially included a sensory score by group (ASD or DD) interaction. If the interaction was not significant in a model, we trimmed the model to include only main effects for ease of interpretation. Covariates in all analyses included service use (therapy and educational services) child gender, age at study start, IQ proxy, household income, mother’s education, and time elapsed between data collection points. All models were run as multiple regression models using SAS software “proc glm” (SAS Institute, 2012).

3. Results

Descriptive results of the key study variables are included in Table 2. Regression diagnostic tests were run with no evidence of outliers or of non-normality in the residuals. Across models, a variety of group interactions and main effects were significant. Where group did moderate the associations of sensory features with adaptive outcomes, the groups’ slopes and their differences are presented in Table 3. In this section, all significant results and non-significant trends are described within each sensory response pattern.

Table 2.

Sample descriptives for key study variables

Study variables—M (SD)		ASD (n = 50)	DD (n = 31)
Sensory scores [possible range 1–5]
	HYPO (parent-report measures)	2.20 (0.57)	1.89 (0.44)
	HYPO (observational measures)	2.27 (0.52)	1.72 (0.46)
	HYPER (parent-report measures)	2.39 (0.48)	2.14 (0.47)
	HYPER (observational measures)	2.27 (0.37)	2.15 (0.33)
	SIRS (parent-report measures)	2.48 (0.49)	2.22 (0.56)
	SIRS (observational measures)	2.42 (0.51)	2.19 (0.54)
VABS scores— 100 (15)
	Daily Living Skills	55.29 (22.50)	54.74 (22.36)
	Socialization	64.02 (16.58)	69.19 (17.62)
	Communication Adaptive Behavior Composite	70.86 (27.34) 59.70 (20.07)	60.45 (20.19) 56.00 (18.41)
Total hours of services between time points
	Therapy services (hours per week)	1.85 (1.22)	1.83 (1.54)
	Educational services (hours per week)	16.93 (12.94)	19.76 (11.95)

Notes. ASD, autism spectrum disorder. DD, other developmental disability. Sensory scores were averaged across measures: parent-report measures were Sensory Experiences Questionnaire Version 3.0 (Baranek, 2009) and Sensory Profile (Dunn, 1999), observational measures were Sensory Processing Assessment (Baranek, 1999b) and Tactile Defensiveness and Discrimination Test—Revised (Baranek, 1998). VABS, Vineland Adaptive Behavior Scales (Sparrow, Balla & Cicchetti, 1984).

Table 3.

Group Slope Estimates for Significant Two-Way Interactions

	Parent Hypo		Parent SIRS				Observed SIRS
Group	ABC	Daily Living	ABC	Daily Living	Communication	Socialization	ABC	Communication
ASD Slope	−5.13 (3.30)	−5.88 (4.65)	−5.69 (4.22)	−7.01 (6.10)	−3.90 (5.23)	−4.55 (4.09)	8.10 (4.20)	9.78 (5.18)
DD Slope	−18.57 (5.89)	−28.06 (9.26)	8.48 (4.55)	9.60 (6.45)	11.22 (5.62)	6.63 (4.40)	−5.33 (4.61)	−5.09 (5.69)
Difference	−13.78 (6.66)	−22.18 (10.37)	14.17 (5.78)	16.60 (7.99)	15.12 (7.15)	11.17 (5.60)	−13.44 (6.24)	−14.88 (7.70)

3.1. Sensory Hyporesponsiveness

The group interactions for parent-reported HYPO were statistically significant in the models predicting both ABC scores, (p = .047) and daily living skills (p = .04). In both cases, the DD group showed a significantly more negative slope than did the ASD group. In other words, we found that as hyporesponsiveness scores increased, ABC and daily living skills scores decreased, and that this effect was more pronounced in the DD group. The group interactions for parent-reported HYPO were not statistically significant for the communication or socialization subscales, nor were any of the group interactions for the observed measures of hyporesponsiveness. There was a significant negative main effect for parent-reported HYPO in the model predicting socialization scores (B = −7.10, p = .02), meaning that higher hyporesponsiveness scores predicted lower socialization scores across groups.

3.2. Sensory Hyperresponsiveness

There was no evidence of statistically significant group interactions for either the parent or observed sensory measures on any of the adaptive behavior outcomes. In our trimmed models, there were significant main effects of parent-reported HYPER in the models predicting both ABC scores (p =0.02) and daily living skills (p < 0.001). Specifically, we found that as parent-reported hyperresponsiveness scores increased, ABC and daily living skills scores decreased across both groups.

3.3. Sensory interests, repetitions, and seeking behaviors

The group interactions for parent-reported SIRS were significant in the models predicting all four adaptive behavior outcome scores (ABC, p = .02; daily living, p = .04; communication, p = .04; socialization, p = .049), and for the observational measure of SIRS on ABC scores (p = .03). There was also a marginal effect for the group interaction for observed SIRS in the model predicting communication scores (p = .06). When the model is trimmed to exclude the interaction, the main effect for observed SIRS is non-significant (p = .67) indicating that if there is an association, it varies across groups. Across all of the parent-reported SIRS models, we observed a similar pattern of interaction, demonstrating that higher parent-reported SIRS scores predicted higher adaptive behavior outcome scores in the DD group and lower adaptive behavior outcome scores in the ASD group. The interaction pattern was reversed for the observed measures; as higher observed SIRS scores predicted higher ABC and communication scores in the ASD group, whereas they predicted lower ABC scores in the DD group.

4. Discussion

This study investigated the longitudinal prediction of three sensory response patterns (hyporesponsiveness, hyperresponsiveness, and sensory interests, repetitions, and seeking behaviors) on subsequent adaptive behaviors (daily living skills, communication, socialization, adaptive behavior composite score) in children with ASD and DD. The amounts of traditional therapy and educational service usage between approximately age 5 and age 9 along with several demographic characteristics were included as covariates in the analyses. Based on conceptual arguments and findings across prior studies (Lane et al., 2010; Liss et al., 2006; Rogers et al., 2003; Watson et al., 2011), we hypothesized that higher levels of sensory features at Time 1, and most notably hyporesponsiveness, would predict lower scores on adaptive behaviors, especially socialization and communication domains, in children in both groups at Time 2. Findings from our analyses indicated that early sensory patterns predicted adaptive behavior outcomes with differential associations by group across particular sensory response patterns as well as the format of measures used (parent reported vs. observational measures).

4.1. Differential Associations between Sensory Response Pattern and Adaptive Behavior Domain

As we hypothesized, higher parent-reported sensory hyporesponsiveness in early childhood years predicted lower overall adaptive behaviors and daily living skills outcomes in later childhood years for both children with ASD and DD. This pattern was also consistent with a negative, non-significant trend for socialization outcomes. Our longitudinal results extend the cross-sectional literature (Ausderau et al., 2016; Watson et al., 2011), supporting the hypothesis that there may be long-term negative implications of sensory hyporesponsiveness on adaptive outcomes. Children with hyporesponsiveness demonstrate diminished responses to specific sensory input in their environments, including adult and peer attempts for social engagement. Thus, children with high levels of hyporesponsiveness early in life may be especially vulnerable to cascading effects on development over time, perhaps due to reduced opportunities to engage meaningfully and learn from their physical environments and social experiences, thereby falling increasingly further behind their peers.

We also observed that higher parent-reported scores of early sensory hyperresponsiveness predicted lower overall adaptive and daily living skills for both children with ASD and DD in later childhood. Children who experience high levels of hyperresponsiveness are acutely aware of specific sensory input in their environments. Thus, the discomfort they experience when engaging with their sensory surroundings may similarly limit the extent to which they engage safely and adaptively with their environments to fully benefit from learning opportunities.

Results in the area of sensory interests, repetitions, and seeking behaviors were less straightforward. In this section we will focus on the parent-reported results because we are less confident about the results from the observed SIRS measure (see following section 4.3). Our results from the parent-reported measures indicated that higher SIRS predicted lower adaptive behavior functioning in the ASD group, suggesting that this sensory pattern may also be impeding the development of adaptive behavior skills for this group. In these instances, it may be that children’s repeated, focused engagement with the sensory aspects of their environments precludes their participation in adaptive learning opportunities. However, the parent-reported measure findings indicated that SIRS may be having the opposite effect for children with DD, suggesting that these sensory features may be serving more of an adaptive function for this group. It is possible that the differential associations we observed between the groups could be related to differences in the nature of the sensory seeking behaviors between groups or differences in perceptions of parents regarding the extent to which SIRS behaviors interfere with versus promote engagement and participation. Our selected assessments specifically measure the frequency of SIRS, but some have suggested that understanding the complexity of these behaviors may shed additional light on the ways in which these behaviors affect daily life (Kirby, Little, Schultz, & Baranek, 2015a). When examining SIRS complexity, Kirby et al. (2015a) observed higher rates of complex behaviors among children with ASD compared to those with DD and TD. Thus, more research is needed in this area to understand possible reasons why SIRS behaviors could be associated with positive adaptive behavior outcomes in the DD group and negative adaptive behavior outcomes in the ASD group.

4.2. Diagnostic Group Differences in ASD compared to DD for Hyporesponsiveness

There were notable differences in the degree of negative association between early hyporesponsiveness in children diagnosed with ASD compared to those with DD. For children with mild early hyporesponsiveness, later adaptive behavior abilities were comparable between groups; however, those in the DD group with high hyporesponsiveness were susceptible to significantly lower Time 2 adaptive behavior abilities than the ASD group. Interestingly, Watson and colleagues (2011) did not find an ASD versus DD diagnostic group interaction for hyporesponsiveness in their cross-sectional study. Perhaps because hyporesponsiveness does not represent a core feature of children with DD as it does for children with ASD (Baranek et al., 2006), it is not similarly targeted in intervention for children with DD compared to their ASD peers, thus increasing the susceptibility of young children with DD who manifest extreme hyporesponsiveness to poorer adaptive outcomes over time. Future research investigating services received by both groups in more detail, including the specific goals of services, is needed. Furthermore, a pattern of sensory hyporesponsiveness (which tends to be less problematic in children with DD as a whole compared to ASD; Baranek et al., 2006) may interact with other co-existing sensory response patterns differently for the ASD group than the DD group.

4.3. Parent Reported vs. Observational Measures

The observed measures of SIRS behaviors in the present study demonstrated opposite findings to the parent-report measures. This lack of congruence in findings using parent-reported measures versus observed measures of sensory behaviors is perplexing but not entirely unexpected. Previous findings in the literature indicate that parent-reported measures and observational measures, although contributing to the same latent construct, may be only modestly correlated—perhaps due to the different contexts in which children are observed by examiners versus caregivers (Baranek et al., 2013). We are cautious about interpreting the results from SIRS observed measures in this study in light of several factors. First, the models with observed sensory measures yielded a much less consistent pattern of results overall than models using parent report measures. A second consideration is that laboratory measures, although more controlled, may be less generalizable to a multitude of naturalistic contexts wherein children may experience a variety of aversive or engaging stimuli in real-world situations that impact adaptive functioning across time. For SIRS behaviors specifically, it is possible that older children are more consciously aware of their unusual mannerisms and may be able to suppress them for short periods in structured testing situations, whereas a wider range of atypicalities may be witnessed by parents across contexts and time.

4.4. Limitations and Future Directions

Certainly, a longitudinal study without an experimental design is limited in causal explanations, but helps to establish precedence of sensory variables as a plausible contributor to adaptive outcomes that is congruent with theoretical arguments (e.g., Baranek et al., 2013). To better determine longitudinal causal relationships, randomized controlled trials are warranted with assignment of interventions and attention to specific dimensional profiles of sensory features (Uljarević, Baranek, Vivanti, Hedley, Hudry & Lane, 2017) or more homogenous subgroups of children (Ausderau et al., 2016; Lane, Molloy & Bishop, 2014). Individuals with ASD are more likely than individuals with DD to show patterns of high hyporesponsiveness concurrent with high hyperresponsiveness (Baranek et al., 2006). Co-occurring sensory response patterns (i.e., interactions between response patterns) were not modeled in the current study; however, it is possible that such extreme co-occurring patterns may exacerbate interference with development of adaptive behaviors over time. Future research accounting for this complexity in light of the interactions reported in this study are warranted.

5. Conclusion

This longitudinal study extends the cross-sectional research in the literature by demonstrating that early sensory features, assessed by both observational and parent-report measures, significantly predict later adaptive behaviors in children with ASD and DD after controlling for services received and demographic covariates. Furthermore, significant results were noted for three sensory response patterns (i.e., hyporesponsiveness, hyperresponsiveness and SIRS), with differential effects based on diagnostic group. Across groups and sensory patterns, we found some evidence that greater frequency of sensory features early in childhood predicted lower adaptive behavior skills later in childhood. Results from this study strongly suggest the need for future research using a developmental perspective to further investigate potential causality and to determine effective services to maximize adaptive behavior outcomes particular to children with ASD and DD, especially those who have high levels of sensory features early in the childhood years.

What this paper adds?

As children age, development of adaptive skills is needed to perform daily life activities that become increasingly complex and demanding. Research indicates that children who exhibit unusual sensory responses to their environment (i.e., hyperresponsiveness, hyporesponsiveness, or intense sensory interests, repetitions, and seeking behaviors) also exhibit increased difficulty in communication, socialization, or completion of daily living tasks. However, research investigating the longitudinal prediction of later adaptive behaviors by sensory features early in childhood is currently lacking. This study expands on existing work by:

• conducting a prospective study examining the extent to which early sensory response patterns (i.e., hyperresponsiveness, hyporesponsiveness, sensory interests, repetitions, and seeking behaviors) predict later adaptive behaviors

• measuring and analyzing each sensory response pattern independently with each adaptive behavior domain

• including diagnostic group as a moderator to deepen understanding of identified associations

• providing evidence from both parent-report and observational measures in longitudinal design

Unique study findings indicate that early sensory features significantly predict later adaptive outcomes, and that these results vary by sensory response pattern. Diagnostic group is associated with some of these differential outcomes.

Sensory Features as Predictors of Adaptive Behaviors: A Comparative Longitudinal Study of Children with Autism Spectrum Disorder and other Developmental Disabilities

Highlights.

• longitudinal study on early sensory features and later adaptive behaviors

• diagnosis and service usage were moderators of analyses

• associations varied, diagnosis and services contributed to differential outcomes