Background
The ability to engage in appropriate social gaze behavior with others is arguably one of the most important skills in the development of human social and communication functioning. Extensive work has determined that maintaining appropriate social gaze is a critical prerequisite for language development, emotion recognition, social engagement, and general learning through joint attention [
1‐
4]. Maintaining social gaze with others not only serves to regulate face-to-face social interactions during early social development, but also facilitates the later coordination of visual attention between individuals or objects of interest [
4,
5]. The development of appropriate social gaze behavior can therefore be considered a critical target for behavioral interventions for children with developmental disabilities.
A subgroup of children with developmental disabilities who are particularly prone to exhibit impairments in social gaze behavior are individuals with fragile X syndrome (FXS). FXS is caused by mutations to the
FMR1 gene at locus 27.3 on the long arm of the X chromosome [
6] which results in excessive methylation of the gene and subsequent reduced or absent fragile X mental retardation protein (FMRP). Characteristic phenotypic features of the disorder include impairments in intellectual functioning [
7] and an increased risk for autistic-like behaviors (e.g., social avoidance, communication impairments, and repetitive behaviors) [
8,
9]. A hallmark behavioral feature of children with FXS is a propensity to avoid social gaze during demanding social situations [
10]. Studies have shown that boys with FXS typically engage in increased social gaze avoidance and social withdrawal during social interactions with unfamiliar people compared to individuals who are more familiar [
10‐
13]. In a naturalistic eye-tracking study, for example, Hall and colleagues [
14] reported that individuals with FXS spent less time looking at an unfamiliar experimenter and exhibited shorter inter-episodes of social gaze when compared to age and IQ-matched controls [
14].
In addition to the characteristic behavioral phenotype of eye gaze avoidance in individuals diagnosed with FXS, data from neuroimaging studies have shown that these individuals exhibit neural circuitry abnormalities in brain regions associated with social behavior. For example, Bruno and colleagues showed that when individuals with FXS were required to look at static photographs of faces in the scanner, neural signaling across multiple brain regions was slower to habituate compared to age and IQ-matched controls [
15]. Another study conducted by Garrett and colleagues demonstrated that individuals with FXS exhibited decreased activation in brain regions known to respond to face and gaze stimuli (i.e., fusiform gyrus and superior temporal sulcus) [
16].
Despite notable behavioral and neural circuitry deficits in social functioning exhibited by individuals with FXS, there is a long-standing debate among practitioners as to whether interventions should be implemented to improve social gaze behavior in FXS. For example, Scharfenaker, O’Conner, Stackhouse, Braden, and Gray [
17] issued the following guidelines to practitioners who implement interventions for individuals with FXS: “Allow avoidant eye gaze. Don’t insist on eye contact” (p.371) and “Position the child so eye contact is not always demanded” (p. 387). Similarly, in their guide for behavioral practitioners working with children with FXS, Morris and colleagues contended that requiring eye contact may result in a number of drawbacks, including “signs of anxiety, more gaze aversion, and possible escalation of problematic behavior” [
18] (p.99). These guidelines may, in part, have resulted from observations that individuals with FXS, particularly males, exhibit elevated levels of physiological arousal (termed “hyperarousal”) when required to complete demanding tasks or to engage in social interactions with others [
8,
9,
12,
19]. Although this association has often been reported in the literature, recent research indicates that hyperarousal in FXS may not necessarily be associated with social interaction. In a recent review of studies conducted on individuals with FXS, for example, Klusek and colleagues [
20] have suggested that the potential increases in physiological arousal that occur in response to social or cognitive stressors in FXS may, in fact, be of similar magnitude to that seen in controls. Rather, these authors suggest that hyperarousal in FXS may be the result of a chronic underlying physiological state rather than a context-dependent phenomenon. Individuals with FXS may therefore be responding to social or cognitive stressors at levels appropriate to the situation [
9,
21,
22].
If this is the case, failing to teach appropriate social gaze behavior to individuals with FXS could ultimately be detrimental to the person’s social development, as well as significantly reducing access to future educational opportunities. For example, several studies have indicated that engaging in frequent social gaze avoidance can negatively impact social interaction skills and communication flow, given that crucial non-verbal gestures and facial expressions that usually aid social interaction will be missed [
23,
24]. Devising interventions that could target appropriate social gaze behavior in FXS, without increasing social gaze avoidance and anxiety, would therefore appear to be particularly important.
One such method that could facilitate the acquisition of social skills in individuals with FXS is discrete trial instruction (DTI)—a standardized teaching procedure that utilizes individualized instruction, well-defined steps, and a consistent rate of training trials to enhance learning [
25,
26]. Indeed, DTI can be advantageous precisely because it involves administering multiple learning opportunities at a relatively high rate—roughly 50–100 per hour [
27]—therefore maximizing the learner’s exposure to the relevant contingencies and ultimately facilitating acquisition of the targeted skill over a relatively short period of time.
DTI has been successfully employed to teach a variety of skills to children diagnosed with autism spectrum disorder (ASD), including imitation [
28], play, vocal, and non-vocal communication, as well as social skills [
29]. Several previous studies have also demonstrated that DTI can be effective for teaching appropriate social gaze behavior to children with ASD or intellectual disabilities [
29‐
31]. In an early study conducted by Foxx [
29], for example, three children with ASD or intellectual disabilities, aged 6 to 8 years, were taught to maintain eye contact with a therapist using DTI. On each trial, the therapist held up an edible item to the eye region and delivered a verbal prompt “Look at me.” If a participant engaged in eye contact, the therapist then delivered the edible item and verbal praise. If the participant failed to make eye contact within 5 s, the therapist implemented an overcorrection procedure by verbally and physically prompting the participant to move his head in one of three directions for a fixed period of time. For all three children, social gaze duration with the therapist improved.
In a study by Cook and colleagues [
30], 21 children with ASD, aged 3 to 12 years old, were each exposed to a progressive training model designed to increase instructional control of social gaze. The model included the following phases: contingent praise only, contingent edibles plus praise, stimulus prompts plus contingent edibles and praise, and contingent video and praise. Results from this study demonstrated that using this procedure allowed the therapists to thin the reinforcement schedule (e.g., from edibles to praise) to some extent in all participants and to maintain learning at or above 80% for 17 of the participants.
Finally, Rapp and colleagues [
31] further modified the protocol described by Cook et al. [
30] to improve social gaze behavior in 15 children with ASD, aged 3 to 7 years. If social gaze was not acquired using the standard progressive model, one or more of the following modifications were made: prompts to sit, altered or skipped phases, altered prompts, or mastery criteria at 70% or higher for five consecutive sessions. Results indicated that social gaze improved for 11 of the 15 participants using the progressive model, but that for 8 participants, one or more procedural changes were required for the children to acquire eye contact. Four participants were unable to acquire social gaze despite modifications to the treatment protocol.
In both the Cook et al. [
30] and Rapp et al. [
31] studies, however, the dependent variable was the latency to engage in social gaze rather than the duration of social gaze. Thus, in both of those studies, it is unclear to what extent the children actually exhibited social gaze behavior, once instructional control had been achieved. To date, only one study has been conducted to examine whether social gaze duration can be improved in children with FXS. Hall, Maynes, and Reiss [
32] employed DTI to teach six boys with FXS, aged 8 to 17 years to engage in increasingly longer durations of social gaze during social interactions with a therapist over a 2-day period. In the Hall et al. study, boys with FXS were asked to look while the therapist held an edible item close to his eye in a series of discrete trials, similar to the procedure utilized by Foxx [
29]. As the trials progressed, a percentile schedule of reinforcement stipulated whether longer durations of social gaze would meet the criteria for reinforcement. For three participants, an overcorrection procedure similar to the procedure employed by Foxx [
29] was required, given that those children did not initially exhibit appropriate social gaze when prompted at the beginning of training. Results from this study demonstrated that although each participant varied in improvement, for 5 of the 6 participants, social gaze duration increased steadily across trials. This suggested that contrary to guidelines issued by some practitioners, behavioral skills training resulted in improved eye contact duration in FXS, at least in the short term.
The present study
In the present study, we refined the teaching paradigm utilized by Hall and colleagues [
32] in a number of ways to facilitate the teaching of appropriate social gaze behavior to boys with FXS. First, to address the potential for elevated autonomic nervous system reactivity (i.e., “hyperarousal”) commonly reported in males with FXS, we paired the DTI procedure with stress- and anxiety-reduction relaxation training. In addition, we used verbal and physical prompts rather than an overcorrection procedure to facilitate the initiation of social gaze and employed contingent verbal praise and a token economy system rather than immediate delivery of an edible item to reinforce appropriate social gaze. To examine the dosage effect of the treatment, we randomized participants to receive the treatment at one of two dose levels:
high or
low. Finally, to examine whether improvements in social gaze duration might generalize to a naturalistic social encounter with an unfamiliar person, participants received a brief social challenge before and after treatment.
We had three research questions:
1.
To what extent would administration of DTI plus relaxation training result in improved social gaze duration in boys in FXS?
2.
To what extent would exposure to a higher dose of treatment result in greater improvements in social gaze duration in boys in FXS?
3.
To what extent would exposure to the treatment result in improved social gaze behavior and decreased physiological arousal during a subsequent naturalistic social encounter?
Discussion
Social gaze is arguably one of the most important skills required for appropriate social interaction. Previous research suggests that engaging in appropriate levels of social gaze during social interactions with others is a critical prerequisite for social engagement [
4]. Accordingly, it is possible that exhibiting increased levels of social avoidance during social interactions may compound delays in social skills. Although social gaze avoidance is a pervasive behavior in boys with FXS, targeted interventions designed to ameliorate this behavior have not been developed. This may, in part, be due to the opinion held by some professionals in the field that social gaze training may not be beneficial for children with FXS and should be discouraged due to concern for potentially increased anxiety and/or problematic behavior in these children [
17,
18,
41].
In this proof of concept study, building on previous research conducted by Hall and colleagues [
32], we evaluated whether a targeted behavioral treatment could be employed to facilitate social gaze behavior in boys with FXS. To date, behavioral interventions designed to ameliorate these intransigent and socially stigmatizing behaviors in FXS have not been evaluated, particularly in the late childhood/early adolescence period when social gaze avoidance may become established in the child’s repertoire. This developmental period can thus present a significant challenge to boys with FXS.
Over the course of treatment, social gaze duration improved significantly for six of the ten boys who received high-dose treatment and for three of the ten boys who received low-dose treatment. Boys who received a greater number of DTI trials therefore benefited more from the treatment. To determine whether the skills learned during the treatment generalized to a naturalistic social setting, we also administered a social challenge in which the participant was required to engage in a conversation with an examiner. To provide an objective measure of social gaze behavior during the social challenge, an eye tracker positioned between the participant and the examiner simultaneously recorded social gaze levels of the participant. Results of this generalization test indicated that social gaze increased to a greater extent in those who had received the high-dose treatment compared to those who had received the low-dose treatment. Physiological arousal levels recorded simultaneously on a heart rate monitor did not appear to be impacted by the treatment.
There are several advantages to the design of the study presented here. First, the intervention utilized is an established and empirically supported behavioral training procedure designed to promote social gaze with a therapist. A further advantage of this study is the use of a percentile schedule, which provided consistent and reliable reinforcement, adapted to the child’s previous performance level. This resulted in a personalized delivery of reinforcement that took into account baseline levels of social gaze and enabled more precise shaping of the participant’s behavior.
There are also several limitations. First, the intervention required two researchers to be present in the room with the participant, a therapist to implement the treatment and another observer to operate the computerized percentile schedule and to record the duration of eye gaze. This may limit the ability for the procedure to be replicated in certain clinical settings. Second, neither the therapist nor the observer was blind to each participant’s group assignment, which may have affected measures of eye gaze duration recorded. Third, the current intervention may need to be adapted to better accommodate the needs of minimally verbal and non-verbal individuals. Furthermore, fatigue experienced by some of the participants over the course of the 2 days of treatment may have resulted in an eventual decreased rate of learning. Anecdotally, it appeared that boys who exhibited the highest levels of social gaze at baseline quickly became fatigued during the intervention and began to exhibit increased numbers of escape behaviors or simply “gave up” trying to look at the therapist. For these participants, it seems likely that more naturalistic forms of social gaze training, such as the methods employed by Carbone et al. [
5] could be used rather than discrete trials.
An additional limitation is that there was no control to examine the effectiveness of relaxation training; rather, all participants regardless of treatment condition received similar amounts of relaxation training. Due to the absence of this control, future research comparing the impact of DTI with and without relaxation training may be helpful. Lastly, the short-term nature of the treatment, and the absence of a parent-training component, may limit the generalizability of the results.
Although DTI has a solid empirical foundation, we acknowledge that there is currently increased emphasis on the promise of naturalistic developmental behavioral interventions, which are still based on the principles of ABA, but are administered within a more naturalistic framework. It is possible that these approaches may have resulted in better generalization and less problem behavior (given that they are believed to be less stressful for the child). A lengthier follow-up study, perhaps over weeks or months, that utilizes alternative more naturalistic ABA intervention methods may therefore be warranted.
Further data could be collected to strengthen the findings of the current study such as using the eye tracker to record social gaze behavior automatically during the intervention [
14]. Interestingly, Miller and colleagues [
42] assessed whether a computer-assisted instruction (CAI) package (pairing visual and vocal stimuli) improved social gaze in three children with ASD. The CAI involved showing the participant a digital picture of a familiar adult accompanied by an audio recording of them saying “look at me!”, while the participant’s social gaze was tracked with an infrared camera. If the participant successfully made eye contact with the face on the computer screen within 5 s, the participant was provided a reinforcing item. Results demonstrated both significant increases in social gaze duration and decreased latency from adult vocalization to making eye contact across participants. These findings may support the future use of CAI to teach social gaze and interpersonal social skills in children with FXS. Future studies could also investigate the extent to which younger children with FXS are amenable to the intervention, before social gaze avoidance becomes established in the child’s repertoire. Studies could also examine whether integrating parent training into the intervention, rather than relying on trained therapists, results in greater generalizability of learned skills.