Autism is characterised by impairments in three main areas of development: social relatedness, communication skills, and the presence of stereotyped behaviour, interests and activities (American Psychiatric Association 2000; Szatmari 2000). Stereotyped behaviours typically associated with autism include hand-flapping, finger flicking, rocking, spinning and self-injury such as head-banging and hand biting (Aarons and Gittens 1999; American Psychiatric Association 2000; Case-Smith and Bryan 1999; Volkmar 1998). Self-stimulation and adopting unconventional postures, such as walking on tiptoes, have also been observed in individuals with autism (Aarons and Gittens 1999; American Psychiatric Association 2000; Ermer and Dunn 1998). One hypothesis for the presence of stereotyped behaviours in autism is that dysfunction in processing sensory information characteristic of the disorder results in the adoption of aberrant behaviours in an attempt to make sense of and regulate stimulation from the environment (Baranek et al. 1997; Nelson 1984; Paluszny 1979).

Sensory processing (SP) refers to the way that sensory information e.g. visual, auditory, vestibular or proprioceptive stimuli is managed in the cerebral cortex and brainstem for the purpose of enabling adaptive responses to the environment and engagement in meaningful daily life activities (Johnson-Ecker and Parham 2000). SP theory suggests that optimal functioning in daily environments requires efficient reception and integration of incoming sensory stimuli. Adaptive behaviour, learning and coordinated movement are considered products of effective sensory integration (Bundy et al. 2002; Kranowitz 1998). Disorders of SP in children are increasingly discussed in the literature (Bundy et al. 2002). Dunn (1997) proposed a model for classifying patterns of dysfunction in SP according to individuals’ behavioural response to stimuli and neurological thresholds, describing four patterns of SP dysfunction: Low Registration, Sensation Seeking, Sensory Sensitivity and Sensation Avoiding.

Unusual responses to sensory stimuli and SP difficulties exhibited by individuals with autism have been widely documented (Baranek et al. 2006; Tomchek and Dunn 2007). Abnormalities have been reported to occur across all sensory domains, including tactile, vestibular, auditory and visual (Harrison and Hare 2004; Rogers 1998) and in the absence of known peripheral dysfunction such as a visual or hearing loss (Baranek 2002). A recent study comparing SP in children with an autism spectrum disorder (ASD; n = 235) to age-matched, typically developing controls (n = 221) reported that 95% of the sample with ASD versus 16.8% of the controls demonstrated some degree of SP difficulty (Tomchek and Dunn 2007). Sensory responses have also been shown to fluctuate such that both hyper- and hypo-responsiveness to sensory stimuli can occur in the same individual (Baranek 2002; Baranek et al. 2006; Greenspan and Wieder 1998; Kranowitz 1998; Volkmar 1998). Further, findings from six studies, which specifically compared the SP patterns of individuals with autism or another pervasive developmental disorder (PDD) with controls, all revealed the presence of significantly different SP profiles for individuals with autism/PDD (Baranek et al. 2006; Ermer and Dunn 1998; Kern et al. 2006; Kientz and Dunn 1997; Miller et al. 2005; Watling et al. 2001). These findings suggest that SP dysfunction is a feature of autism.

Despite the abundance of descriptive literature and anecdotal reports documenting unusual responses to sensory stimuli in this population, the nature of SP difficulties in autism and the relationship to the core functional and behavioural disturbances of the disorder remains poorly understood (Baranek 2002). In particular, there is inconsistency in reports regarding the nature of SP patterns within autistic disorder (AD) (Rogers and Ozonoff 2005; Schaaf and Miller 2005). Further, the relationship between SP and behavioural difficulties in autism has only recently received attention from researchers and warrants further investigation (Miller et al. 2005; Rogers et al. 2003).

Patterns of SP dysfunction may be related to core features of the disorder and the development of unusual behaviours. They may also give insight into prognosis, therefore having significant implications for early diagnosis and intervention (Baranek 2002; Ermer and Dunn 1998). Improved understanding of the way in which children with autism process sensory information is required (Tomchek and Dunn 2007). The purpose of this preliminary study was to: (1) describe SP patterns of children who have AD, and (2) explore the relationship between SP patterns and emotional, social and behavioural responsiveness in this group.

Methods

Participants

Children diagnosed with AD were recruited from the Early Intervention Research Program (EIRP) for children with autism at Flinders University, South Australia, Australia. Children are eligible to take part in the EIRP if they are under the age of 5 years and meet DSM-IV criteria for AD. Inclusion criteria ensured that only children who had received a diagnosis of AD by a multi-disciplinary team using standard diagnostic tools were included. Standard diagnostic tools used to diagnose participants included the Autism Diagnostic Interview-Revised (ADI-R; Rutter et al. 2003), the Childhood Autism Rating Scale (CARS; Schopler et al. 1988), CHAT (Baron-Cohen et al. 2000), the Autism Detection in Early Childhood tool (ADEC; Young 2006), the Adaptive Behaviour Assessment System (2nd ed.; ABAS-II; Harrison and Oakland 2003) and the Diagnostic and Statistical Manual (4th ed.; DSM-IV; American Psychiatric Association 2000). Children were eligible to take part in this study if they were either on the waiting list, currently enrolled, or had completed the EIRP. Children who met the criteria for AD but who presented with a co-morbid diagnosis that impaired intellectual functioning, i.e., additional disorders such as a chromosomal abnormality, were excluded from the study.

Materials

Short Sensory Profile (SSP):

The SSP is a 38-item parent questionnaire designed to measure behaviours associated with abnormal responses to sensory stimuli in children aged 3–10 years (McIntosh et al. 1999). The SSP is administered as a questionnaire in which the caregiver rates the child’s typical responses to tactile, vestibular, auditory and visual stimuli on a five point Likert scale ranging from ‘never’ responds in this manner to ‘always’ responds in this manner. Item scores cluster meaningfully into patterns of SP such as Tactile Sensitivity and Low Energy/Weak, enabling sensory sensitivity and performance to be interpreted in each of these areas. Higher scores relate to typical performance expected from children whereas lower scores demonstrate that a definite difference is present. The SSP was derived from a longer caregiver questionnaire, the Sensory Profile (Dunn 1999) but is accepted as the most appropriate version to use in research protocols (Dunn 1999; McIntosh et al. 1999). Good reliability of the SSP has been demonstrated in a previous study and acceptable internal and discriminative validity has also been established (Dunn 1999). A separate study also demonstrated adequate construct validity (Miller et al. 2005). The SSP takes approximately 10 min to complete.

Vineland Adaptive Behaviour Scales (VABS), Interview Edition:

The VABS is a semi-structured interview administered to caregivers to assess the current adaptive behaviour of the child, defined as the development and application of abilities required for the attainment of personal independence and social competence (Sparrow et al. 1984). The VABS assesses adaptive behaviour by obtaining standard scores in the domains of communication, daily living, socialisation and motor skills. An optional Maladaptive Behaviour domain measures undesirable behaviours that may hinder the adaptive skills of the subject. Items in each domain are scored from 0 to 2, with decreasing scores indicating skills/behaviours that are sometimes or never performed. The raw scores are then converted to standard scores. The sum of standard scores in each domain yields an adaptive behaviour composite score reflecting the overall ability of the subject to live independently. All domains, with the exception of motor skills, were analysed within the current study. The VABS is routinely administered to caregivers of children enrolled in the EIRP.

Developmental Behaviour Checklist—Parent (DBC-P):

The DBC-P is a 96-item parent checklist developed to assess a range of behavioural and emotional problems in children and young people (4–18 years) with an intellectual disability (Einfeld and Tonge 2002). The total score provides a threshold for clinically significant behavioural and emotional problems and sub-scales assess individuals in the areas of disruptive and anti-social behaviour, self-absorbed behaviour, communication disturbances, anxiety, social relatedness and autism specific behaviours (DBC-ASA). The parent edition of the DBC (DBC-P) reflects behaviours exhibited by the child over a 6-month period. The DBC-P is routinely administered to caregivers of children enrolled in the EIRP.

Procedures

Ethics

Ethics approval was gained from the Division of Health Sciences Human Research Ethics Committee at the University of South Australia and Flinders University Social and Behavioural Research Ethics Committee.

Recruitment

Information packages were mailed to individuals who were on the wait list, enrolled in or who had completed the EIRP for children with autism at Flinders University and who met the inclusion criteria. Informed consent was obtained and data collection occurred at the EIRP House, Flinders University in a single face-to-face session, except for one participant who completed the caregiver questionnaire at home and mailed it to the researcher.

Data Collection and Analysis

The SSP was administered to parents according to the guidelines outlined in the SSP manual (Dunn 1999). DBC-P and VABS data were obtained by gathering each participant’s most recent data set for these assessments from their EIRP records. As SP and behavioural data were gathered at separate times, counterbalancing was not considered to be relevant. Although the most recent data sets for most participants were dated within 12 months of the current study, data for six participants were from assessments conducted between 1 and 2 years prior and two participants’ results were of assessments administered more than two years prior. Given the relatively small sample size within this pilot study, analysis of results excluding the eight participants’ older data sets was not considered appropriate. SSP results and findings from behavioural assessments were analysed initially using descriptive statistics. Pearson’s Correlation Coefficient method was then used to correlate data from the SSP, the DBC-P and the VABS.

Results

Participants

Twenty-two children with AD were recruited to the study and all had completed the EIRP at the time of data collection. Participants were aged between 33 and 101 months (2 year 9 months–8 year 5 months), with a mean age of 64.86 months (5 year 5 months; 20.70 SD). Eighteen participants were male, resulting in an approximate ratio of 5:1 male to female which is consistent with gender ratios currently reported in the literature for autism (American Psychiatric Association 2000; Dempsey and Foreman 2001).

Patterns of Sensory Processing Performance

For the 22 children recruited to the study, SSP scores were classified as an overall score, into three performance categories and across seven areas of SP. The mean overall score for the SSP was 133.36 (SD = 20.70) and scores ranged between 83 and 160. Eighteen participants (82%) demonstrated either a probable or definite difference in SP (probable, n = 6; definite, n = 12). Figure 1 shows the distribution of participant scores across SSP sections. Most participants demonstrated a definite difference in the Under-responsive/Seeks Sensation and Auditory Filtering sections (n = 15, 68%). Twelve (80%) of these participants exhibited a definite difference in both sections. Typical performance was demonstrated by 16 participants (73%) in Movement Sensitivity and by 12 (55%) in Visual/Auditory Sensitivity. Ten (45%) participants demonstrated typical performance in both Movement Sensitivity and Visual/Auditory Sensitivity sections. Remaining participants in both the Movement Sensitivity and Visual/Auditory Sensitivity sections were equally divided between the Definite Difference and Probable Difference performance categories.

Fig. 1
figure 1

Sensory processing performance by SSP section. Key: TS = Tactile Sensitivity; AF = Auditory Filtering; TSS = Taste/Smell Sensitivity; LEW = Low Energy/Weak; MS = Movement Sensitivity; VAS = Visual/Auditory Sensitivity; USS = Under-responsive/Seeks Sensation; DD = definite difference; PD = probable difference; TP = typical performance

Full size image

Other notable SP performance patterns included (a) an equal number of participants showed a definite difference and typical performance in Low Energy/Weak, with no participants demonstrating probable difference in this section, and (b) performance by the sample on Tactile Sensitivity and Taste/Smell Sensitivity was relatively equally distributed between definite, probable and typical performance.

Table 1 displays the results of the correlational analyses.

Table 1 Correlational analysis of SP performance patterns
Full size table

As seen in Table 1 moderate, significant associations were observed between Movement Sensitivity and Tactile Sensitivity, Tactile Sensitivity and Visual/Auditory Sensitivity and Taste/Smell Sensitivity and Low Energy/Weak. Moderate correlations, which were significant at the 0.01 level included Movement Sensitivity and Visual/Auditory Sensitivity, Taste/Smell Sensitivity and Under-responsive/Seeks Sensation and Under-responsive/Seeks Sensation and Low Energy/Weak. Two clusters of association between areas of SP performance were revealed in this analysis—Cluster One: Visual/Auditory, Tactile and Movement Sensitivity and Cluster Two: Low/Energy Weak, Taste/Smell Sensitivity and Under-responsive/Seeks Sensation.

Behavioural Responsiveness

From Table 2 it is evident that most participants showed low levels of performance for each of the VABS domains included in this study as well as significant scores for maladaptive behaviour however no significant behavioural difficulties as measured by the DBC-P were evident.

Table 2 Participants’ scores on VABS and DBC-P
Full size table

Relationship Between SP Performance Patterns and Behavioural Responsiveness

Pearson’s correlation analyses of SSP and VABS and DBC-P scores showed several significant relationships, as displayed in Table 3.

Table 3 Analyses between sensory and behavioural data
Full size table

Total SSP scores and total DBC-P scores were found to be strongly negatively associated as were total SSP and VABS Maladaptive Behaviour domain scores. A moderate, positive correlation was calculated between total SSP and VABS Daily Living Skills scores. Moderate, negative correlations were found between the VABS Maladaptive Behaviour domain and Movement Sensitivity, Under-responsive/Seeks Sensation, Auditory Filtering and Low Energy/ Weak. Total DBC-P scores were also moderately, negatively correlated with several SSP sections, including Underresponsive/Seeks Sensation, Auditory Filtering, Low Energy/Weak and Visual/Auditory Sensitivity. These results indicate that, in general, higher VABS Maladaptive Behaviour domain and DBC-P total scores were associated with lower SSP scores.

Discussion

In this study, overall SSP results showed that most participants in this sample (82%) exhibited some degree of SP difficulty. These results are consistent with the findings of recent studies that report the prevalence of SP dysfunction in children with autism to range between 69% and 95% (Baranek et al. 2006; Schaff and Benevides 2007; Tomchek and Dunn 2007). Despite the lack of a control group, normative data reported in the SSP suggests these SP difficulties are of greater magnitude than might be expected in the typically developing population. In order to assess the uniqueness of these difficulties to autism, these data need to be compared with data collected from other disability groups. Results from the current study were also based on the chronological age of children only and previous research supports a stronger relationship between SP and developmental age than SP and chronological age (Baranek et al. 2006). SSP results from the current study showed that the majority of participants had marked impairment in several areas of SP while demonstrating typical responsiveness in others, indicating the presence of specific patterns of SP functioning.

Areas of Impairment in SP

Most participants showed a definite difference in the Under-responsive/Seeks Sensation and Auditory Filtering sections with only one participant in each of these sections demonstrating typical performance. Kern (2002) suggested that under-responding to sensory stimuli in individuals with autism may be explained by the depression of sensory afferents in the cerebellum, resulting in inconsistent sensory modulation. In a study by Watling et al. (2001), Sensation Seeking was identified as a section in which children with a PDD demonstrated significantly lower scores from the typically developing group, congruent with the findings of the current study. However, conflicting results were observed by Ermer and Dunn (1998), who reported a low incidence of atypical behaviours in the Sensory Seeking category for the AD/PDD group. Behaviours associated with sensory seeking, such as inappropriate smelling, licking and rubbing, may account for stereotyped and repetitive behaviours observed in this population.

Auditory Filtering has also been shown to be an area of difficulty for children with AD and other PDDs in previous studies. Findings from studies by Ermer and Dunn (1998) and Watling et al. (2001) showed a high frequency of behaviours in the Inattention/Distractibility category of the Sensory Profile, an equivalent category to the Auditory Filtering section within the SSP, congruent with current findings. When a child has difficulties with auditory filtering, he or she may be either hyper-responsive i.e., sensitive to sounds, or hypo-responsive, i.e., oblivious to sounds (Dunn 1999). Children with auditory filtering difficulties may appear to be distracted and inattentive, particularly in busy environments, which contain numerous stimuli such as the school environment.

Areas of Typical Performance in SP

Most participants showed typical performance in the Movement Sensitivity and Visual/Auditory Sensitivity sections of the SSP. It is noteworthy that most participants who demonstrated typical performance in these sections showed typical performance in both of these sections, suggesting that these areas may be associated in some way. A Pearson’s correlation coefficient of .615 (p = .002) confirmed that a significant, moderate relationship existed between these sections. Typical responsiveness in movement sensitivity was also observed for a group of children with AD/PDD in another study (Watling et al. 2001), whilst a further study yielded contrary results (Ermer and Dunn 1998). These conflicting results highlight the need for further investigation of patterns of typical responsiveness as this may assist clinicians to maximise areas of relative strength in SP and consequently support learning and adaptation for individuals with AD. Individuals with AD often show superior skills and repetitive behaviours in visual tasks such as building blocks, pattern recognition and puzzles, which may be associated with the high level of performance observed in the Visual/Auditory Sensitivity section in the current study.

Areas of Varied Response Within SP

In the current study, several SSP sections elicited a mixed response from participants, with a greater spread of scores occurring in these sections. The most pronounced example was observed within the Low Energy/Weak section in which exactly half of the participants demonstrated a definite difference and half showed typical performance, illustrating distinct patterns of both hypo- and hyper-responsiveness within the same section. Taste/Smell Sensitivity was another area in the current study in which participants demonstrated varying performance, with almost even numbers of participants receiving scores for definite difference and typical performance. Large variance in the Taste/Smell Sensitivity and Low Energy/Weak sections was also noted in a study by Watling et al. (2001) but contrasting results were observed in a study by Ermer and Dunn (1998) where children with AD/PDD exhibited a high frequency of behaviours within the SSP Taste/Smell Sensitivity section.

Clusters of SP Responsiveness

Several clusters of SP responsiveness were noted in the current study, one of which wholly replicated, and the other partially replicated, findings from a previous study (Miller et al. 2005). The first cluster was observed between Tactile Sensitivity, Movement Sensitivity and Visual/Auditory Sensitivity, which were also detected by Miller et al. who titled this cluster ‘Sensory-Over-Responsivity’. Another cluster of SP symptoms observed by Miller et al. titled ‘Sensory Under-Responsivity’ included Taste/Smell Sensitivity, Low Energy/Weak and Visual/Auditory Sensitivity to a lesser degree. Results from the current study partially support this cluster, with a moderate correlation observed between Low Energy/Weak and Taste/Smell Sensitivity but with the Under-responsive/Seeks Sensation section replacing Visual/Auditory Sensitivity.

SP and Behavioural Responsiveness

Various significant relationships were found between SP patterns and behavioural responsiveness. Strong, negative correlations between SSP scores and total DBC-P and VABS Maladaptive Behaviour domain scores indicated that in general, poor SP ability was associated with higher levels of behavioural and/or emotional problems for this sample. A moderate correlation found between total SSP and VABS Daily Living Skills scores, indicated that within this sample, poorer SP ability was associated with decreased functioning in the area of daily living skills. Other authors have hypothesised that SP difficulties limit participation in activities of daily living, work and leisure activities (Schaaf and Miller 2005). Both the VABS Maladaptive Behaviour domain and total DBC scores showed moderate correlations with the following SSP sections: Under-responsive/Seeks Sensation, Auditory Filtering and Low Energy/Weak. The DBC Autism subset (DBC-ASA), which measures autistic behaviour, was also found to be moderately correlated with these three SSP sections. In a study by Miller et al. (2005), sensory seeking behaviours were found to be strongly and consistently related to a higher incidence of repetitive behaviours. Repetitive behaviours are likely to constitute much of the maladaptive behaviour seen in the current sample. If an individual demonstrates under-responsive or sensory seeking behaviours, more stimulation is required in order to regulate and process sensory information received from the environment, which in turn may manifest in the form of repetitive behaviours.

Conclusion

In summary, several areas of marked impairment, typical responsiveness and varied responsiveness were observed in this study, suggesting the presence of distinct SP patterns for children with AD. In general, participants demonstrated SP difficulties in auditory filtering and a high incidence of sensation seeking behaviours, areas, which mainly focused on under-responsiveness in relation to movement and vestibular input. Other distinctive SP patterns in this sample included typical responsiveness in the areas of Visual/Auditory and Movement Sensitivity and varied responses for Low Energy/Weak and Taste/Smell Sensitivity SSP sections. Of the two distinct clusters observed which either wholly or partly replicated groupings found by Miller et al. (2005), most participants in the current study demonstrated typical performance for all SSP sections within the Sensory-Over-Responsivity cluster but showed a definite difference for sections within the Sensory-Under-Responsivity cluster.

Noteworthy findings observed between SP and behavioural data include significant moderate correlations between both maladaptive behaviour and emotional/behavioural problems, and the following SSP sections: Under-responsive/Seeks Sensation, Auditory Filtering and Low Energy/Weak. Replication of this study with a larger sample size is warranted to further clarify the observed relationships and to develop hypotheses regarding the impact of SP on behavioural responsiveness in autism. Testing of these hypotheses will then assist in the identification of specific SP and behavioural typologies within autism that will enable more targeted approaches to intervention and therefore improved allocation of resources for this population.