Reference no: EM133317464

Assignment: Heterogeneity in the autism endophenotype and treatment response

Autism spectrum disorders (ASDs) are a group of neurodevelopmental disorders characterized by core deficits in three domains: social interaction, communication, and repetitive or stereotypic behavior, with variable degree of impairment among individuals and impact on affected families.

Is has been suggested that ASD is one of the most familial of psychiatric disorders. Twin studies have demonstrated a high heritability for ASDs around 80-90% (Rutter, 2000). Nonetheless, clinical, and epidemiologic studies suggest that gene penetrance and expression may be influenced, in some cases strongly, by environmental factors (Eapen, 2011). Environmental risk factors perhaps play a role via complex genetic-epigenetic-environmental factor interactions, but no specific exposures with significant population effects are known. A number of endogenous biomarkers associated with autism risk have been investigated, and these may help identify significant biologic pathways that, in turn, will aid in the discovery of specific genes and exposures (Newschaffer et al., 2006). Genetic analyses indicate genetic heterogeneity with considerable overlap with other disorders such as intellectual disability and attention deficit hyperactivity disorder, confirming the documented clinical heterogeneity in symptom expression (Charman et al., 2011) and suggesting simultaneous genetic variations in multiple genes (Dawson et al., 2002; Eapen, 2011). Due to the evident heterogeneity, ASDs are considered as a spectrum of conditions that affect individuals differently, although distinct phenotypic expressions are masked by the limitations of diagnostic symptom representations (Eapen, 2012; Eapen et al., 2013).

Endophenotype research holds considerable promise for the study of gene-brain/cognition-behavior pathways for developmental disorders (Viding and Blakemore, 2006). In the field of ASDs various atypical neurocognitive profiles and neurophysiological alterations that have been obtained from neuroimaging, eye tracking, and electrophysiological studies are reported. In early years of life, common behavioral expressions of low-level impairments of social attention and reciprocity are reported: reduced preference and attention to persons and other social stimuli reduced respond to vocal approaching, poor verbal imitation, poor establishing of eye contact, and recognition of emotions. Neurocognitive approaching explored the cognitive-behavioral phenotype and a number of cognitive models of ASD have been proposed over time: the theory of mind-blindness, weak central coherence and executive functions. To date, few distinct behavioral subtypes have been identified but poorly replicated. Generally, although the total number of factors varied, the majority of studies reported at least one social-communication factor and at least one distinct non-social factor comprising repetitive behaviors and restricted interests. Taken as a whole, the literature suggests that the clinical presentation of individuals on the autism spectrum varies with respect to level of functioning and comorbid disorders. Specifically, the proposed neurocognitive models do not seem to be consistently related to measures of symptom severity and social competence and are appearing on other disorders such as ADHD and intellectual disability evidenced only modest correlations between socialization, communication and repetitive/restricted interests and behaviors, suggesting multiple-deficit accounts considering the developmental and dynamic aspects of individual profiles. Thus, do neither account for all characteristic symptoms of ASDs nor are necessarily specific to ASDs. As yet, there is no single theory, which integrates all characteristics of ASD.

The heterogeneity of ASD may not only underlie the insufficiency of single-cause neurocognitive models in explaining the triadic autism phenotype but may also underlie the fast variability in treatment response that is documented in ASD treatment efficacy studies. Meta-analyses and systematic reviews have generally concluded that Early Intensive Behavioral Interventions (EIBIs) based on the principles of applied behavior analysis appear to be the most effective treatment for ASD to date. Children following this approach demonstrate significant improvements in the areas of autism severity; cognitive, language, and adaptive functioning, and aberrant behaviors. However, although EIBI resulted in improved outcomes for children with ASD applying analysis strategies at a comparison group level, there was a fast variability detected when analysis was conducted at an individual within-group level (Howlin et al., 2009). Therefore, research pointing at factors and methods allowing for treatment individualization is warranted. The crucial question in EIBI research has shifted from general effectiveness toward understanding why outcomes vary across different children and for which children is EIBI most and least effective (Kasari, 2002). Such studies should shed light on which children benefit most from which interventions and the intensity and length of treatment necessary to effect a change. The heterogeneity and developmental nature of ASDs make it unlikely that one specific treatment model or its specific implementation strategy will work for any one child throughout his cognitive and social development. Research points clearly toward the inadequacy of a "one-size-fits-all" approach favoring a single treatment program for all areas of learning in all contexts for children with ASD (Stahmer et al., 2011). Teasing out the active ingredients of effective treatment appears to be fundamental to the refinement of strategies and procedures that work best for specific settings, subgroups of children, or providers (Kasari, 2002). This requires an understanding of pre-treatment child and family characteristics as well as specific intervention strategies and delivery formats associated with differential treatment response rates. However, available research provides only limited information on individual outcomes and moderator or mediator variables of the varying developmental trajectory. Rates of less than20%of early intervention articles that measured possible treatment outcome moderators have been established. Nonetheless, available evidence from meta-analyses indicates a number of predictive pre-treatment child characteristics and specific treatment factors associated with response to treatment. These include larger gains in overall IQ facilitated by higher treatment intensity, applied supervisor training, and applied parent training, larger gains in adaptive functioning facilitated by higher treatment intensity, the inclusion of parent training, and higher pre-treatment language skills as well as longer treatment duration, and larder gains in language skills facilitated by longer treatment duration. Furthermore, varying strength of predictors has been detected when meta-regression analysis was controlled for subgroup differences between specific EIBI delivery models. Specifically, treatment intensity and supervisor training contributed to the efficacy of staff-directed EIBI programs in producing improved IQ and adaptive outcomes, while in contrast, child pre-treatment IQ, language and adaptive skills gain predictive strength the higher the extent of parent inclusion EIBI programs. Given the heterogeneity of ASD, EIBI programs, treatment response rates as well as of the strength of outcome predictors, it is likely that a client-centered personalized approach increases the efficacy of a program. This requires multi-disciplinary research that is accounting for endophenotypic profiles in explaining subgroup differences in treatment response as well as how behavioral intervention techniques and delivery formats address each of these differences. Such a stance would improve implications for personalized intervention planning as a one-treatment-fits-all approach leads to mixed treatment response, given the many profiles of autism and the different developmental trajectories individuals may follow. Thus, the following paper discusses a set of issues critical to treatment individualization and thus crucial for any research approach aiming to detangle factors related to heterogeneous treatment response rates and differing developmental trajectories. The set of critical issues - namely that programs that begin earlier, are more intensive, more comprehensive, and require parent inclusion to lead to better response rates while specific risk patterns of child characteristics lead to non-response - is not exclusive but has been chosen due to the widespread belief in their accuracy. The final goal of this paper is to extract implications from this set of critical issues that allow modifying treatment components with respect to developmental and individual needs, thus enhancing the efficacy of EIBI programs for a wide range of behavioral profiles present in children with ASDs and lately to indicate utile areas were to integrate clinical behavioral and basic neurodevelopmental research, from an integrative umbrella point of view.

2. Critical issues important to treatment response and individualization

A set of critical issues that are widely accepted in their accuracy and appear recurring in the literature - namely that intervention programs that begin earlier, are more intensive, more comprehensive, and require parent inclusion led to better response rates while specific risk patterns of child characteristics lead to non-response are critically discussed. This is achieved by applying 5 principles from the conceptual biobehavioral framework of efficient early interventions for children with developmental disabilities introduced by Ramey and Ramey (1998).

2.1. Principle of developmental timing

Interventions that begin earlier in development afford greater benefits to the participant than do those that begin later (Ramey and Ramey, 1998). The hypothesis that the age at the start of treatment is a major variable for determining outcomes vast implications for the field. It has been assumed that the critical age of entering behavioral intervention is prior to four years. To date, there is no study available testing this hypothesis ad-hoc with a methodologically sound design. Anyways, a range of intervention studies leads to conflicting results suggesting that this variable may not be as important for the outcome as thought. While younger age at intake predicted a better outcome in some studies, others found the absence of such influence. However, it has been assumed that age-at-intake effects have been detected due to the wider age range of these samples. Furthermore, some studies detecting age at intake effects based their results on school placement as outcome measures and target mastery. These outcome measures are undoubtedly important criteria for progress evaluation and inclusion but do not account either for successful participation in community settings nor for valid changes in the autism core symptoms and related abilities. Meta-analyses provide the possibility to account for age at intake as a moderating variable including sample information from original studies that did not conduct such data analysis. Nevertheless, four meta-analyses that did conduct such moderator-analyses did not provide evidence for age at intake effects on outcome. Makrygianni and Reed (2010) commented that besides the absence of a significant association of age at intake and treatment outcome, there is noteworthy increase of variability in outcome when children become older, which could be attributed to various interfering factors. The results of our treatment efficiency study on the cross-setting EIBI program specified that although significant associations of age at intake with child outcome was lacking at between-group level, some relations were identified at a within-group level of analysis (Strauss et al., 2012). Younger age at intake was associated with larger gains in cognitive and language outcome measures, while such association for adaptive behavior abilities approaches zero. Interestingly, an age-at-intake effect was absent in the eclectic comparison group which demonstrated a more narrow age range than the behavioral intervention group which had an approximately 13 month wider age range. The initial methodological flaw of having not fully age matched comparison groups secondarily enabled to provide support for the Eikeseth et al. (2007) assumption of underlying age range group differences accounting for age at intake effects on outcome as well as for the general call for developmental appropriateness in skill teaching. We assume that increase in age range goes in line with the increase in heterogeneity of autism symptom constellations and related skill deficits. Therefore, age at intake effects may not represent a predictor of general EIBI efficacy rather than indicate developmental timing at skill level questioning how to tailor program components in respect to the individual's age and ability considering developmental precursors of a targeted skill.

Lately, neurodevelopmental research on brain plasticity in autism is examining sensitive periods in development during which behavioral treatment is likely to have its greatest impact. Sensitive periods in brain development are defined by an increased flexibility of the circuitry in a young, developing brain, represented by the development of more connections than needed (Katz and Shatz, 1996), by the increased activity of cellular mechanisms forming and eliminating connections (Knudsen, 2004) and by the absence of fully shaped and activated neural circuit patterns (Hensch, 2005). This implies that by actively shaping early experience one can actively shape the architecture of the developing brain before it is fully stabilized. In the field of autism, the hypothesis of deprivation of critical early experience-driven input that result from failure to pay attention to social stimuli is derived from the evidence of face recognition (Klin et al., 1999; Osterling and Dawson, 1994) and speech processing impairments (Klin, 1992; Kuhl et al., 2005, 2013). Early experience refers to the interaction of a child with its environment. Nevertheless, it has been suggested that the simple exposure to experience, e.g., language does not necessarily facilitate the development of brain circuitry specialized for its perception (Kuhl, 2007) but needs to affect emotional arousal via the rewarding function of reinforcement (Dawson et al., 2002; Kampe et al., 2001). Thus, early intervention promoting attention to social stimuli and engagement in social interaction during sensitive periods in brain development may mitigate effects of ASD on brain development. Dawson et al. (2012) offered ESDM intervention emphasizing interpersonal exchange, social attention and shared engagement leading to improved autism severity, IQ, language and adaptive skills outcome. Nevertheless, ERP and EEG measurements were correlated solely with levels of social behavior at outcome but not with measurements of autism symptoms, IQ, language and adaptive behaviors. These results, point to the potential of early intervention to alter the course of brain development in young children with ASD but imply a clear specificity of intervention input and developmental output. It remains unclear if interventions focusing on a social interactive context for speech perception, or more sophisticated cognitions (planning actions, determining emotional responses) lead to comparable brain development as well as it remains unclear which are sensitive periods facilitating higher-level brain circuits.

In absence of direct evidence for the "the earlier the better" hypothesis and the fact that many children (particularly those who are more able) do not receive a diagnosis before aged four years, an alternative hypothesis of "better late than never" has been assumed (Howlin, 2003). In the presence of implications toward efficacy of EIBI for different abilities at different ages (Strauss et al., 2013) and the assumption that different kinds of experiences are critical at different ages for optimal brain development (Greenough et al., 1987) a "what and when?" approach to the problem may be more appropriate.

2.2. Principle of program intensity

Programs that are more intensive produce larger effects than do less intensive interventions. Children and Parents that are participating more actively and regularly are the ones who show more developmental progress (Ramey and Ramey, 1998). The hypothesis that more intensive interventions - in terms of treatment hours - will yield greater gains for children with ASDs has been shifted toward the controversy of a point of diminishing[1]returns where the child does not improve significantly from more treatment (Matson and Smith, 2008). Surprisingly, to date an insufficient amount of research has evaluated the role of an EIBIs intensity considering that already the seminal paper by Lovaas (1987) included a controlled comparison of treatment intensity and introduced intensity as a main factor impacting treatment outcomes. In that study, a high-intensity group (40 h a week) improved significantly more in cognitive skills than the low-intensity group (10 h a week). Since then, very few studies have directly addressed that issue, with a major part of studies that did not find intensity to effect EIBI efficacy. Nevertheless, Reed et al. (2007) provided a direct comparison of high- and low-intensity interventions basically replicating the Lovaas finding with children receiving EIBI for approximately 30 h a week making significantly greater developmental gains than those receiving 12 h a week. Other studies focused on predictive models and found treatment intensity to interact with other possible predictors of developmental trajectories. Granpeesheh et al. (2009) established a quadratic relationship between treatment intensity and skill mastery outcomes for different age groups. The youngest age group (2-5 years) was highly responsive already at low levels of intensity, the middle age group (5-7 years) responded best to treatment at high levels of intensity, while children older than seven years showed stable response rates unaffected by the amount of treatment exposure. Osborne et al. (2008) established a possible interaction of intensity and parental stress affecting outcome. In general, greater skill improvements are achieved during more time-intensive interventions (16-40 h a week) than those with lower time-input (1-15 h a week), but only when parenting stress levels are low. Specifically, children of highly stressed parents that were enrolled in high time-input programs fell significantly below mean improvement rates. A recent meta-analysis demonstrated insufficient reporting practice regarding the information on treatment intensity (Strauss et al., 2013), specifying that five studies out of21reportedanapproximate range of treatment hours, four studies missed reporting treatment hours for control groups, one study did not report treatment intensity data at all. Despite difficulty in obtaining clear estimates of treatment intensity, results from meta-regressions indicated that the improvement of the child's abilities is generally affected by the intensity of the program. Nevertheless, several specifications have been elaborated. Behavioral EIPs have been found to be very effective at around 25 h per week or more (Makrygianni and Reed, 2010). This effect was strongest for adaptive behavior outcomes. Viruès-Ortega (2010) found increasing effect sizes related to increased programs intensity solely for adaptive behavior outcomes, while for intellectual functioning exhaustion of such intensity-intervention effects has been shown. This examination has been deepened by Strauss et al. (2013) that found the influence of treatment intensity to change with the extent of parent inclusion. Increase in intensity of staff-provided programs increased the program's effectiveness to result in IQ score change, while high-intensive programs that include parents in treatment provision are most effective for adaptive behavior change. Thus, currently available results on the treatment intensity provide support that it is not simply the programs intensity measured in treatment hours or duration that produce larger effects, but moreover it is about the active involvement of the participating children and their treatment providers that needs to be studied more specifically. It should be noted that studying intervention intensity measured by quantity does not reflect all aspects of intensity (e.g., different settings, different providers, quality of teaching input and supervision). The current knowledge implies that treatment-specific and child-specific variables are not independent but will influence one another (Strauss et al., 2013). Thus, it does not seem that a simple increase in treatment hours will linearly involve an increase in behavioral outcome. To date it is not clear, how to establish for each individual child an optimal range of hours per week that will deliver the best developmental achievement, without exhausting children and staff alike. Nevertheless,the basic assumption of EIBI is that developmental progress is a product of learning. There is poor acknowledgment in the recent research literature of how the measured quantity of time is used in terms of quality or time spent by parent generalizing gains made in one-to-one therapy. Evaluating the parent inclusion in the cross[1]setting EIBI program, we found the child's developmental progress measured as maintenance and generalization of mastered targets as well as decrease in challenging behaviors being directly influenced by increase in target difficulty appropriate to the child's improving skill profile, by parent provided sessions at home in addition to center-based one-to-one treatment, and indirectly by staff and parent fidelity (Strauss et al., 2012). Thus, child's skill trajectory depended on appropriate individualized program planning, active parent involvement, and adherence to the treatment plan developed via shared-decision making between primary treatment providers and involved caregivers. Similarly, we found that active parent involvement and a child-oriented teaching style fostering self-initiation in sessions targeting play and social interaction led to more functional play, peer proximity, and social interaction overtures from the target child than a highly structured adult-directed teaching style in sessions where parents were excluded from active participation (Strauss et al., 2014). These results imply that the conceptualization of a treatment's intensity could be more appropriate by including measures such as target intensity, the intensity of parent inclusion, the child's skill preparedness, spontaneity, and self-initiation. Similar implications are provided from research on language development in children with autism, demonstrating that not the quantity but the type of utterances (responsive and contingent verbal input) facilitate positive language outcomes in children with ASD (Haebig et al., 2013; McDuffie and Yoder, 2010; Siller and Sigman, 2008). Specifically, follow-in comments to stimuli or activities that are already in the child's current focus of attention were associated with language development, whereas redirecting or out-of-focus comments were negatively associated with later language production. This underlines the importance of quality not quantity enhanced activities where parent and child are jointly focused on the same referent.

Our assumption that simple increase in treatment intensity (quantity) does not necessarily produce more developmental gains without considering how stimuli are provided during intervention (quality), led support from the neurodevelopmental assumption that simple exposure to specific stimuli does not necessarily relate to the development of specialized brain circuitry (Kuhl, 2007). Thus a stronger inclusion of neurodevelopmental and clinical research is demanded. As, early experience refers to the interaction of a child with its environment, learning results from an interaction of the child's capacities and the environmental input. Within the child, organization of the brain and connectivity of specialized areas directly impacts processing and recalling, which in turns shapes learning and development of skills. Behavioral interventions apply specific strategies directing the stimuli input in order to establish motivation and initiation of functional behaviors, more recently focusing on natural occurring social reinforcement rather than reward-based reinforcement. Nevertheless, individuals with ASDs have been characterized by low responsiveness to social rewards such as facial expressions, verbal cues and gestures. Likewise, a decrease in performance has been associated with the use social than non-social rewards independently from the number of rewards provided (Demurie et al., 2011; Geurts et al., 2004). Neurodevelopmental research clearly demonstrated brain-behavior underpinnings of lacking engagement between the treatment provider and the involved child by poor reward-function of social cues (Kohls et al., 2012) as well as by deficient attention-shift mechanism (Akshoomoff et al., 2002), deficit imitation mechanisms (Williams et al., 2001), and deficit internal simulation mechanisms (Oberman and Ramachandran, 2007). Kohls et al. (2012) point on the evidence for disrupted reward-seeking tendencies in social contexts, interfering with incentive-based motivation and learning, are likely caused by a dysfunction in the social motivation "wanting" circuitry (dopaminergic-oxytocinergic). This results in difficulty in attending and processing facial and behavioral cues, undermining the child's capacity to engage with and learn from others.

It seems to be crucial to clarify the brain-behavior underpinnings of aberrant social motivation and the effectiveness of essential reward systems applied in behavioral interventions by integrating behavioral and neurodevelopmental findings. Though, neurodevelopmental studies rely on off-context task designs that lack two fundamental components of everyday face-to-face interactions: contingent responding and joint attention. The recent development of a live face-to-face interaction design during fMRI may bridge the gap between previous research investigating the neural basis and behavioral aspects of social interaction (Redcay et al., 2010) by providing a task design that includes the complexity of dynamic, multimodal social interactions

2.3. Principle of individual differences in program benefits Some children show greater benefits from participation in early intervention than do other children. Thus far, the individual differences appear to be related to aspect of the child's initial risk condition (Ramey and Ramey, 1998). The idea that individuals respond differently to the same program leads to the consequence that program planning tailored to individual needs is demanded. Previous research has indicated a number of child's initial risk conditions that are possibly associated with different treatment response. Generally, higher pre-treatment IQ , language and social communication, adaptive skills, imitation skills, and joint attention lead to greater developmental gains. Some studies, however, lead to inconsistent results confirming the heterogeneity in ASDs characteristics and thus the heterogeneity in treatment response, with no impact of pre-treatment IQ but of IQ change during treatment, with higher pre-treatment autism symptom severity predicting positive gains while others found less severe children making more progress. Most studies based their prediction models on mean difference scores in order to detect such differences that are largely determined by the improvements made in the comparison group. Some recent meta-analyses used mean change scores providing information on the magnitude of improvements achieved in one specific group under one specific condition. They conclude that behavioral EIBIs are equally effective for children with very low or medium intellectual abilities as well as for both verbal and non-verbal children. Furthermore, Makrygianni and Reed (2010) and Strauss et al. (2013) found a relationship of intake and outcome language and adaptive abilities. Language abilities of the children improve more when the children start with high adaptive behavioral skills and vice versa. Surprisingly, despite the clear call for the inclusion of measurement of ASD core symptoms and challenging behaviors, most EIBI studies failed to address these dependent variables (Matson, 2007). This is kind of disappointing, as children with severe ASD exhibit significantly more and higher levels of challenging behaviors (Matsonet al., 2008) with stereotypical behaviors being the most reported (Jang et al., 2011). We underlined the importance of including measures of autism severity and challenging behaviors in effectiveness studies in our evaluation of the cross-setting EIBI program (Strauss et al., 2012). Indeed, autism severity, specifically social communication and social interaction deficits, was found the best predictor for child performance during acquisition and generalization tasks, while the amount of challenging behaviors was found best predictors for incorrect responding or failure to respond during teaching. Interestingly, subtype analysis confirmed that stereotypic behaviors did mostly impede the child's performance in both, parent and staff provided sessions. Lastly, higher child performance and less challenging behaviors were significantly associated with better outcome in language skills, adaptive functioning and mental development state. Anyhow, little information is provided about who does well in which treatment and why, as no study referred to a different treatment. Studies identifying predictors of response to one specific treatment are important, but are limited to provide conclusions on how specific intervention techniques address each of those pre-treatment characteristics. Such systematic investigation of treatment-behavior interactions is crucial as different behaviors may best approached via different treatment protocols. Thus, although claims are made on the need of treatment individualization (Stahmer et al., 2011), to date, no treatment or family factor as well as their interaction has been clearly identified that allow to tailor where child characteristics demand an individualized treatment.

2.4. Principle of program breadth and flexibility Interventions that provide more comprehensive services and use multiple routes to enhance children's development generally have larger effects than do interventions that are narrow in focus (Ramey and Ramey, 1998). Comprehensive EIBIs for ASDs differ from specific interventions in the width of targets to be addressed, including core deficits in autism, language, social, cognition, daily living, academic, and play skills (Rogers and Vismara, 2008). Likewise specific interventions, comprehensive EIBIs may differ in service delivery parameters such as intensity, settings, and providers (Kasari, 2002). Nevertheless, the common feature of comprehensive EIBIs in ASDs is the combination of a broad program of behavioral principles with a target curriculum that addresses all areas of functioning (Gould et al., 2011). This perspective bases on the idea that superiority claims of one strategy over another are ultimately obsolete as heterogeneity in outcome is a stable finding within and across coexisting strategies (Schreibman and Anderson, 2001). Thus, the issue is not which form of behavioral treatment is better, but rather, which form of treatment is best for a specific child at a specific time. A decision toward the appropriateness of a teaching procedure is intertwined with the skill to be taught. Nevertheless, recommendations toward the constitution of a comprehensive curriculum are missing (Gould et al., 2011) and a simultaneous use of more than one curricular programs by clinicians has been documented, indicating shortcomings in single available curricula (Love et al., 2009). The earliest systematic application of learning theory in autism has been the discrete trial training (DTT) that is addressing deficient skill very directly in developmental order of building blocks toward higher-level skills. In fact, most comprehensive programs and manuals distinct between basic, intermediate and advanced skill repertoires (Fava and Strauss, 2011; Fava et al., 2012; Hayward et al., 2009; Maurice et al., 1996). Skill teaching and program planning is coordinated within each skill domain, thus the introduction of new acquiring skills proceed as soon as pre-requisite target behavior has been achieved. Nevertheless, it is a recurrent finding that comprehensive EIBIs are very efficient in teaching functional communication and other cognitive outcomes while adaptive and socialization behaviors are less easily addressed (Fava et al., 2011; Strauss et al., 2012). Meta-analyses provide a similar pattern with significantly smaller pooled effect sizes for adaptive behaviors than for language skills and cognitive functioning (Makrygianni and Reed, 2010; Strauss et al., 2013; Viruès-Ortega, 2010). The hypothesis of pivotal skills accounts for the recurrent lack of emergence of social-communicative and social interaction skills, which are skills that are correlated to the emergence of others skills and whose delay interferes the development of higher-order skills (Koegel et al., 2001). Thus, coordinating skill teaching and program planning across domains rather than exclusively within domains is crucial as each-domain facilitates the progress of another. Regarding the area of social communication, it has been hypothesized that joint attention, imitation and language might be part of a shared social-communicative representational system that is progressing in specialization and differentiation (Charman et al., 2000). This hypothesis let supports studies confirming inter-relations of increasing joint attention, imitation and language skills (Whalen et al., 2006; Ingersoll and Schreibman, 2006). Toth et al. (2006) specified that Proto declarative joint attention and immediate imitation are precursors for referential speech, whereas toy play and deferred imitation are precursors for social-communicative behaviors. Nevertheless, language training did not result in increased joint attention or play behaviors, suggesting that targeting early social-communicative behaviors may lead to changes in higher-order developmental behaviors but not vice versa. It has been proposed that joint attention, play, imitation, and language might be part of a shared social-communicative representational system, which is progressing in specialization and differentiation (Charman et al., 2000). Furthermore, an atypical developmental sequence of emergence of social-communicative skills has been documented, with autistic children following object- and action-related nonsocial cues while typical children intentionally refer primarily to the social cues in social-communication learning (Carpenter et al., 2002). This atypical order of emergence of social-communicative skills in autistic children is in line with the previously reported documentation of reduced preference to social stimuli (Klin et al., 2009; Nadig et al., 2007; Osterling et al., 2002;Werner and Dawson, 2005) and the consequential poor reward function of social cues (Demurie et al., 2011; Geurts et al., 2004; Kohls et al., 2012). This has implications on teaching. Discrete trial teaching applies mainly non-social rewards in order to establish a social consequence, with children being requested to look at the adult and to point at the preferred object in order to have access at the later one. Nevertheless, the social consequence (looking at the adult and sharing attention) as a mutual affective exchange is rarely accompanying an object-related exchange when being taught this way. Thus, in consequence children taught joint attention exclusively via non-social rewards are less likely to attend and engage in social-communicative exchanges. Embedding social interactions into child preferred non-social reinforcement has been shown to increase child positive affect and child initiated, coordinated engagement in social-communicative behaviors (Vernon et al., 2012). This outcome supports the need for the inclusion of complementary teaching formats. Naturalistic behavioral approaches applying less artificial stimuli and reinforce protocols (Koegel et al., 2001; McGee et al., 1999). These techniques emphasize the increase in the child's choice, varying task sequencing, interspersing acquisition with mastered tasks and the broadening of shaping criteria in order to increase the child's motivation, responsivity to multiple cues and self-initiation of targeted skills. Indeed, as response to the documented lack of significant development of socialization skills in the cross-setting EIBI model (Fava et al., 2011; Strauss et al., 2012)the authors adapted the specific play and social engagement setting and teaching format. In the comparison study of highly structured adult-directed and flexible child-oriented approaching toward teaching a clear benefit from embedding social interactions as reinforcers in a child-oriented approach to teach play behaviors was demonstrated (Strauss et al., 2014). The child-oriented approach led to significantly higher-order functional play in peer proximity as well as to more social-communication overtures toward peers compared to a highly structured adult-directed approach relying on DTT teaching and non-social rewards. Particularly, regression analysis revealed that such achievement was mainly predicted by activities that facilitate the child initiation of social engagement and by less direct prompting of an interfering adult. The neurodevelopmental markers of readiness to change, developmental sequences of emerging skills and neural activity in response to specific teaching formats are lacking but is warranted in order to discern the underlying mechanisms of skill development. Moreover, this might provide the possibility to identify differential markers of targeted behaviors other than for basic social-communication behaviors (such as stereotypies, higher-order pragmatic language skills, and executive functions) useful in individualized treatment planning, such as to decide when to move from a more directive intervention to a more naturalistic, inclusive context. Ultimately, the simple provision of a comprehensive program including a comprehensive skill curriculum does not necessarily lead to larger effects than within a narrow skill program. Moreover, it seems that without coordinated skill teaching and program planning across skill domains, considering the individual readiness to actively engage in target behaviors and positively respond to the chosen teaching and reinforcing strategies, the possible multiple routes to enhance the child's development may not be effectively exhausted or adapted when the transfer is necessary due to child progress.

2.5. Principle of direct provision of learning experience

Children receiving interventions that provide direct experiences show larger and more enduring benefits that do children in programs that rely on intermediary routes to change in child's competencies (e.g., parent training only) (Ramey and Ramey, 1998). Service delivery of EIBIs has been programmed in mainly three different forms, those who are center-based with trained staff and/or parents providing treatment to children with autism, those who are home-based with trained staff and/or parents providing treatment and seeking to enhance the children's everyday learning opportunities, and those who combine center-based and home-based formats. In any case, parent training and their consecutive inclusion in treatment provision has been scheduled since the early days of EIBI treatment evaluation (Lovaas, 1987). Since that a wide range of methods toward effective parent training and parent-mediated intervention has been reported (Matson et al., 2009; McConachie and Diggle, 2006). Parent training is conducted in a wide variety of methods, from group to individual training, home and center-based training using a multitude of tools such as manuals, curriculums, video training, or live instructions. Likewise, varying formats of parent inclusion in treatment provision have been reported - ranging from collaborative staff-parent teams to fully self-directed intervention. Matson et al. (2009) suggested that parent training needs to be conducted in a combination of intensive theoretical and practical teaching in center-based setting. A description of inclusive parent training and combined staff- and parent-mediated treatment provision is provided by Fava et al. (2012) that accounted particularly for treatment fidelity issues Direct comparisons of interventions that are delivered and implemented via face-to-face (professional and parent) or self-directed modalities (the parent alone, with or without supervision) are not in the focus of research, yet. Systematic reviews conclude that parent training and parent-mediated interventions are positively related to child's social communication outcomes and parent child interactions (Keen et al., 2010; McConachie and Diggle, 2006). Nevertheless, clear estimates on the benefit of including parents in center-based treatment provision or fully parent-implemented EIBIs for children with ASDs were actually lacking as positive outcomes were only found on parent report measures. To date, the only study that is systematically evaluating the differential effectiveness of EIBI delivery formats (staff only provision, inclusive staff and parent provision, parent only provision) is provided by Strauss et al. (2013). Using a meta-analytic approach, results from single and pooled effect size calculation suggest that EIBI programs including parents in skill generalization in the home setting in addition to staff-directed center-based intervention led to overall higher effect sizes in intellectual, language and adaptive outcomes. This result does not simply imply that parent inclusion increases EIBI program effectiveness. In contrast, it has been estimated that fully parent-implemented treatment led to the smallest effect sizes in all three outcome measures. Itis noteworthy that the pooled samples within all three delivery formats did not differ significantly in age at intake, pre-treatment language, and adaptive skills, treatment intensity, or treatment duration. This result supports the notion that the context in which parent inclusion is programmed and evaluated must be fully considered. Parents do not only need initial training but also ongoing support as their children develop (McConachie and Diggle, 2006). Factors that may reduce parent inclusion efficiency are suggested such as extensive demands on parents, infrequent training from consultants, reliance on therapists who may have little experience with ASD treatment and background with parent supportive skills, as well as high staff turnover (Smith et al., 2000b). In general, although there is an extensive variety of EIBIs demanding parent inclusion in treatment provision and the need to implement behavior target schedules outside of regularly implemented intervention hours, very few studies provide research data on the effects of systematical parent involvement on both children and the parent themselves (Makrygianni and Reed, 2010). Indeed, Strauss et al. (2013) documented the lack of assessment of the parent's mastery in implemented techniques in most studies. Five out of nine studies claiming complementary staff and parent treatment provision provided indirect evidence for the staff's treatment fidelity, none for the parent's procedural adherence. Three out of five studies evaluating fully parent-mediated treatment programs provided indirect evidence for adherence to the treatment protocol. Nevertheless, none of these studies did actually report data directly. Thus, whether the parents applied the procedures correctly is unknown. Few implications can be drawn from the literature. Intensive parent training combined with intensive theoretical and practical teaching in a center-based setting that aims to establish a staff-parent collaborative team is effective in ensuring parent treatment fidelity (Fava et al., 2011; Strauss et al., 2012). Nevertheless, detailed analysis evidenced a particular difficulty for parents to achieve high fidelity in progress monitoring and in strategies aimed to facilitate the child's engagement in play activities in the natural environment (Strauss et al., 2012). Furthermore, it seems fundamental to establish work collaboration between staff and parent that is based on active inclusion in professional planning and monitoring processes. It has been shown that ineffective staff-parent collaboration can contradict treatment planning and decrease a program's effectiveness (Strauss et al., 2012). Previous research attributed the efficacy of parent inclusion to parental stress levels. Osborne et al. (2008) found that clinical EIBI programs, especially those with higher time-input, are less effective when parenting stress levels are high. Moreover, parental stress decreases in parents who are providing low-intensity treatment, but increases in high-intensive treatment provision (Brookman-Frazee, 2004; Keen et al., 2010). This let to the conclusions that parent involvement potentially increases parental stress as demands on parents increases, impeding positive child outcomes. Nonetheless, parent stress attributed to negative child outcomes in staff-provided interventions as well (Shine and Perry, 2010). Strauss et al. (2012) replicated this finding with high parental stress affecting child outcomes in both parent-mediated EIBI and staff-provided eclectic comparison intervention. Child outcomes that imply a reciprocal interaction between parent and child (expressive language, adaptive behaviors and autism core symptoms) were negatively influenced by pre-treatment parental stress, while cognitive abilities were largely unaffected. Strauss et al. (2012) aimed detangle to what extent the parent inclusion in treatment provision accounts for an appropriate implementation of teaching strategies and as such predicts or counteracts the facilitation of child's behavior improvement across clinical and community setting. Regression analysis revealed that the perception of having a difficult child impeded improvements in autism symptom severity, while high parenting distress let to positive language development, with parents that worked more intensively on the child's daily development facilitating language improvements. Detailed analysis indicated that it is less about the parental stress itself rather than inefficient staff-parent collaborations. It has been shown that high parental stress combined with low staff treatment fidelity compromises decision-making toward appropriate behavior targets. That is when modification in target schedules is done due to parental stress regardless of the child's actual skill. Such inappropriate target scheduling resulted in significant more challenging behaviors during sessions and thus to lower child overall performance. It is not clear yet in which way different perceptions between staff and parent toward both child and parent characteristics, as well as progress variables counteract efficient treatment implementation and how shared decision-making might be facilitated in order to contribute to positive parent and child progress. Solish and Perry (2008) demonstrated lack of correspondence between staff and parent perceptions of rating of involvement, self-efficiency in conducting treatment, autism specific knowledge and parental stress. This is an important result, as it has been shown that the successful involvement of parents is solely predicted by these parent variables particularly treatment self-efficiency. Concordant perception between staff and parent of belief in Intensive Behavior Interventions and the child's progress through the intervention did not predict successful parent involvement in treatment provision. Ultimately, the simple inclusion of parents in treatment provision does not necessarily lead to increase in child's benefit from a specific EIBI program. Research suggests that successful EIBI implementation needs intense parent-professional relationships in order to ensure that treatment gains are maximized. These results underline the need to actively involve parents in all components of the selection and planning of individualized treatment plans, to tailor parent training according to identified parent needs; and to foster shared decision-making during supervision and ongoing child progress monitoring 3. Conclusions and implications It is our conviction that cross-discipline collaboration targeting critical issues important to detangle the high variability in treatment response will provide novel assumptions in future research that might be the key to useful treatment individualization and prevent excessive research repetition without being considerable replications as it is in available treatment efficiency studies. Implications drawn from the literature and the own research experiences of the authors were introduced for five (not exclusive) critical issues to treatment efficacy. We assume the unique perspective that, indeed early identification and treatment begin is crucial for positive developmental trajectories in children with ASD. But, as ASD is a lifelong condition one may keep in mind that different kinds of experiences are critical at different ages for optimal brain development. Thus, as there is a vast lack of research focusing on school-aged children or adolescents, research toward efficacy of EIBI at different ages assuming a "what and when" rather than "the earlier the better" approach is warranted. Furthermore, conceptualizing treatment intensity as quality variable allows for deepened understanding how active involvement of children and treatment providers may be measured, achieved and related to outcome. The current focus on responder profiles defined by child characteristics is pivotal for determining the autism phenotype. Nevertheless, poor insight in treatment techniques-behavior profiles interaction is limiting the current clinical practice of in tailoring treatment to individual needs at different stages of developmental trajectory. The systematic integration of treatment and provider characteristics helps to address such treatment limitations. Research on treatment individualization enables treatment providers systematically to enhance the individual's readiness to engage in target behaviors by coordinated and collaborative program planning across comprehensive skills curricula and treatment packages. This may offer the opportunity to exhaust the possible multiple routes to enhance the child's development, adapt programs successfully when transfer is necessary due to child progress, and to enhance treatment adherence in involved treatment providers, leading to long-lasting behavior change. Collaborative program planning and integration, especially at the level of skill domains and teaching strategies, may upgrade parent-professional relationships as to successful treatment tailoring according to child characteristics demands parent training and family support tailoring according to parent needs, when parents are included in treatment provision and transfer of positive behavior outcome into all important community settings is aimed for. We propose some specific implications for the future research in EIBI program effectiveness that may provide a step forward to an umbrella-like conceptualization of autism research incorporating cross-discipline collaboration toward treatment individualization across treatment packages, skill curricula, child and family needs:

1. The need to strengthen efficient research that is integrating knowledge from the disciplines of genetic, neurological, and behavioral science.

2. The need to link research of sensitive periods in brain development and brain plasticity to clinical treatment progress data as well as the need to go beyond the application of experience-expectant mechanisms to treatment focusing of enhances the experience in the very early life. There is a need to involve research in brain development in later developmental periods by applying experience-dependent mechanisms that are modifying existing circuitry and are unique to the environment not the age, as well.

3. The need to involve components of everyday dynamic social interactions into neurodevelopmental studies that so far rely on off-context task designs. One step forwards is provided by the recent development of a live face-to-face interaction design during fMRI. Further methodological improvements and clinical applications of these techniques are warranted.

4. The need to increase research on treatment individualization across skill domains, treatment packages, and providers that is considering both, the individual child and family needs.

5. The need to broaden the spectrum of clinical study samples reported by the inclusion of study samples of children on the autism spectrum with a wide range of comorbid psychopathologies, by the inclusion of low-functioning autistic individuals and by the inclusion of a wide range of age groups and developmental stages.

6. The need to deepen insight in processes that may ensure successful parent-professional relationships and strengthen treatment adherence by research focusing on the advantages and limitations of shared-decision making that may involve disparities in staff- and parent-reported outcomes and treatment priorities, exact procedural fidelity as well as experienced self-efficiency, and the consequential unmet child and parent needs

Question: Can you summarize the article?