Faster eye movements in children with autism spectrum disorder

Early identification of autism spectrum disorder based on machine learning with eye-tracking data

BACKGROUND

Early identification of autism spectrum disorder (ASD) improves long-term outcomes, yet significant diagnostic delays persist.

METHODS

A retrospective cohort of 449 children (ASD: 246, typically developing [TD]: 203) was used for model development. Eye-movement data were collected from the participants watching videos that featured eye-tracking paradigms for assessing social and non-social cognition. Five machine learning algorithms, namely random forest, support vector machine, logistic regression, artificial neural network, and extreme gradient boosting, were trained to classify children with ASD and TD. The best-performing algorithm was selected to build the final model which was further evaluated in a prospective cohort of 80 children. The Shapley values interpreted important eye-tracking features.

RESULTS

Random forest outperformed other algorithms during model development and achieved an area under the curve of 0.849 (< 3 years: 0.832, ≥ 3 years: 0.868) on the external validation set. Of the ten most important eye-tracking features, three measured social cognition, and the rest were related to non-social cognition. A deterioration in model performance was observed using only the social or non-social cognition-related eye-tracking features.

LIMITATIONS

The sample size of this study, although larger than that of existing studies of ASD based on eye-tracking data, was still relatively small compared to the number of features.

CONCLUSIONS

Machine learning models based on eye-tracking data have the potential to be cost- and time-efficient digital tools for the early identification of ASD. Eye-tracking phenotypes related to social and non-social cognition play an important role in distinguishing children with ASD from TD children.

Identifying Autism Gaze Patterns in Five-Second Data Records

One of the most challenging problems when diagnosing autism spectrum disorder (ASD) is the need for long sets of data. Collecting data during such long periods is challenging, particularly when dealing with children. This challenge motivates the investigation of possible classifiers of ASD that do not need such long data sets. In this paper, we use eye-tracking data sets covering only 5 s and introduce one metric able to distinguish between ASD and typically developed (TD) gaze patterns based on such short time-series and compare it with two benchmarks, one using the traditional eye-tracking metrics and one state-of-the-art AI classifier. Although the data can only track possible disorders in visual attention and our approach is not a substitute to medical diagnosis, we find that our newly introduced metric can achieve an accuracy of 93% in classifying eye gaze trajectories from children with ASD surpassing both benchmarks while needing fewer data. The classification accuracy of our method, using a 5 s data series, performs better than the standard metrics in eye-tracking and is at the level of the best AI benchmarks, even when these are trained with longer time series. We also discuss the advantages and limitations of our method in comparison with the state of the art: besides needing a low amount of data, this method is a simple, understandable, and straightforward criterion to apply, which often contrasts with "black box" AI methods.

No evidence for differential saccadic adaptation in children and adults with an autism spectrum diagnosis

Background

Altered patterns of eye-movements during scene exploration, and atypical gaze preferences in social settings, have long been noted as features of the Autism phenotype. While these are typically attributed to differences in social engagement and interests (e.g., preferences for inanimate objects over face stimuli), there are also reports of differential saccade measures to non-social stimuli, raising the possibility that fundamental differences in visuo-sensorimotor processing may be at play. Here, we tested the plasticity of the eye-movement system using a classic saccade-adaptation paradigm to assess whether individuals with ASD make typical adjustments to their eye-movements in response to experimentally introduced errors. Saccade adaptation can be measured in infants as young as 10 months, raising the possibility that such measures could be useful as early neuro-markers of ASD risk.

Methods

Saccade amplitudes were measured while children and adults with ASD (N = 41) and age-matched typically developing (TD) individuals (N = 68) made rapid eye-movements to peripherally presented targets. During adaptation trials, the target was relocated from 20-degrees to 15-degrees from fixation once a saccade to the original target location was initiated, a manipulation that leads to systematic reduction in saccade amplitudes in typical observers.

Results

Neither children nor adults with ASD showed any differences relative to TD peers in their abilities to appropriately adapt saccades in the face of persistently introduced errors.

Conclusion

Of the three studies to date of saccade adaptation in ASD, none have shown deficits in saccade adaptation that are sufficient to generalize to the whole or a subgroup of the ASD population. Unlike prior studies, we found no evidence for a slower adaptation rate during the early adaptation phase, and no of evidence greater variance of saccade amplitudes in ASD. In post hoc analysis, there was evidence for larger primary saccades to non-adapted targets, a finding requiring replication in future work.

Visual-spatial and visuomotor functioning in adults with neurofibromatosis type 1

BACKGROUND

Neurofibromatosis type 1 (NF1) is a neurodevelopmental genetic disorder associated with visual-spatial and visuomotor deficits, which have not been studied well in adults with NF1.

METHODS

In 22 adults with NF1 and 31 controls, visuomotor functioning was assessed by measuring eye latency, hand latency and hand accuracy during visuomotor tasks. Visual-spatial functioning was assessed by measuring eye movement responses during the Visual Threshold Task.

RESULTS

The NF1 group had a significantly shorter eye latency than the control group and was less accurate in their hand movements during specific visuomotor tasks. The groups showed no differences in eye movement responses during the Visual Threshold Task and in hand latency during the visuomotor tasks.

CONCLUSIONS

In contrast to studies in children with NF1, we found no alterations in visual-spatial information processing in adults. Impairments in eye latency and hand accuracy during specific visuomotor tasks may indicate deficits in visuomotor functioning in adults with NF1.

Visual search in neurodevelopmental disorders: evidence towards a continuum of impairment

Disorders with neurodevelopmental aetiology such as Attention-Deficit/Hyperactivity Disorder (ADHD), Autism Spectrum Disorder (ASD) and Schizophrenia share commonalities at many levels of investigation despite phenotypic differences. Evidence of genetic overlap has led to the concept of a continuum of neurodevelopmental impairment along which these disorders can be positioned in aetiological, pathophysiological and developmental features. This concept requires their simultaneous comparison at different levels, which has not been accomplished so far. Given that cognitive impairments are core to the pathophysiology of these disorders, we provide for the first time differentiated head-to-head comparisons in a complex cognitive function, visual search, decomposing the task with eye movement-based process analyses. N = 103 late-adolescents with schizophrenia, ADHD, ASD and healthy controls took a serial visual search task, while their eye movements were recorded. Patients with schizophrenia presented the greatest level of impairment across different phases of search, followed by patients with ADHD, who shared with patients with schizophrenia elevated intra-subject variability in the pre-search stage. ASD was the least impaired group, but similar to schizophrenia in post-search processes and to schizophrenia and ADHD in pre-search processes and fixation duration while scanning the items. Importantly, the profiles of deviancy from controls were highly correlated between all three clinical groups, in line with the continuum idea. Findings suggest the existence of one common neurodevelopmental continuum of performance for the three disorders, while quantitative differences appear in the level of impairment. Given the relevance of cognitive impairments in these three disorders, we argue in favour of overlapping pathophysiological mechanisms.

Computer-Aided Screening of Autism Spectrum Disorder: Eye-Tracking Study Using Data Visualization and Deep Learning

Background

The early diagnosis of autism spectrum disorder (ASD) is highly desirable but remains a challenging task, which requires a set of cognitive tests and hours of clinical examinations. In addition, variations of such symptoms exist, which can make the identification of ASD even more difficult. Although diagnosis tests are largely developed by experts, they are still subject to human bias. In this respect, computer-assisted technologies can play a key role in supporting the screening process.

Objective

This paper follows on the path of using eye tracking as an integrated part of screening assessment in ASD based on the characteristic elements of the eye gaze. This study adds to the mounting efforts in using eye tracking technology to support the process of ASD screening

Methods

The proposed approach basically aims to integrate eye tracking with visualization and machine learning. A group of 59 school-aged participants took part in the study. The participants were invited to watch a set of age-appropriate photographs and videos related to social cognition. Initially, eye-tracking scanpaths were transformed into a visual representation as a set of images. Subsequently, a convolutional neural network was trained to perform the image classification task.

Results

The experimental results demonstrated that the visual representation could simplify the diagnostic task and also attained high accuracy. Specifically, the convolutional neural network model could achieve a promising classification accuracy. This largely suggests that visualizations could successfully encode the information of gaze motion and its underlying dynamics. Further, we explored possible correlations between the autism severity and the dynamics of eye movement based on the maximal information coefficient. The findings primarily show that the combination of eye tracking, visualization, and machine learning have strong potential in developing an objective tool to assist in the screening of ASD.

Conclusions

Broadly speaking, the approach we propose could be transferable to screening for other disorders, particularly neurodevelopmental disorders.

Random and Short-Term Excessive Eye Movement in Children with Autism During Face-to-Face Conversation

This study investigated the oculomotor performance in children with autism spectrum disorder (ASD) during a face-to-face conversation. A head mounted eye tracker recorded the eye movements in 20 children with ASD and 23 children with typical development (TD). Group comparisons were conducted on the randomness and the quantity of eye movement. The amount of time needed to reveal group difference was also examined. Results showed that the randomness of eye movement was significantly higher at all examined time durations, and the amount of eye movement was significantly greater within 3 s in the ASD group. These findings demonstrated an atypical pattern of oculomotor dynamics in children ASD, which might facilitate the objective identification of ASD during daily social interaction.

Basic oculomotor function is similar in young children with ASD and typically developing controls

A variety of eye tracking studies have demonstrated that young children with ASD gaze at images and movies of social interactions differently than typically developing children. These findings have supported the hypothesis that gaze behavior differences are generated by a weaker preference for social stimuli in ASD children. The hypothesis assumes that gaze differences are not caused by abnormalities in oculomotor function including saccade frequency and kinematics. Previous studies of oculomotor function have mostly been performed with school-age children, adolescents, and adults using visual search, anti-saccade, and gap saccade tasks that are less suitable for young pre-school children. Here, we examined oculomotor function in 144 children (90 with ASD and 54 controls), 1-10-years-old, as they watched two animated movies interleaved with the presentation of multiple salient stimuli that elicited saccades-to-targets. The results revealed that the number of fixations, fixation duration, number of saccades, saccade duration, saccade accuracy, and saccade latency did not differ significantly across groups. Minor initial differences in saccade peak velocity were not supported by analysis with a linear mixed model. These findings suggest that most children with ASD exhibit similar oculomotor function to that of controls, when performing saccades-to-targets or freely viewing child-friendly movies. This suggests that previously reported gaze abnormalities in children with ASD are not due to underlying oculomotor deficiencies. LAY SUMMARY: This study demonstrates that children with ASD perform similar eye movements to those of controls when freely observing movies or making eye movements to targets. Similar results were apparent across groups in the number of eye movements, their accuracy, duration, and other measures that assess eye movement control. These findings are important for interpreting previously reported differences in gaze behavior of children with ASD, which are likely due to atypical social preferences rather than impaired control of eye movements.

Visual Sensory Experiences From the Viewpoint of Autistic Adults

Although previous research has investigated altered sensory reactivity in autistic individuals, there has been no specific focus on visual sensory experiences, particularly in adults. Using qualitative methods, this study aimed to characterize autistic visual sensory symptoms, contextualize their impact and document any associated coping strategies. A total of 18 autistic adults took part in four focus groups which involved questions around visual experiences, the impact of these on daily life, and strategies for their reduction. Transcripts of each session were thematically analyzed allocating six key themes. Participants described a range of visual hypersensitivities, including to light, motion, patterns and particular colors, which contributed to distraction and were frequently part of a wider multisensory issue. Such experiences had significant negative impacts on personal wellbeing and daily life with participants describing fatigue, stress and hindrances on day-to-day activities (e.g., travel and social activities). However, the degree of understanding that participants had about their visual experiences influenced their emotional response, with greater understanding reducing concern. Participants employed a variety of coping strategies to overcome visual sensory experiences but with varied success. Discussions also highlighted that there may be a poor public understanding of sensory issues in autism affecting how well autistic individuals are able manage their sensory symptoms. In summary, autistic adults expressed significant concern about their visual experiences and there is a need to improve understanding of visual experiences on a personal and public level as well as for developing potential support.

Autism Detection in Children by Combined Use of Gaze Preference and the M-CHAT-R in a Resource-Scarce Setting

Most children with autism spectrum disorder (ASD), in resource-limited settings (RLS), are diagnosed after the age of four. Our work confirmed and extended results of Pierce that eye tracking could discriminate between typically developing (TD) children and those with ASD. We demonstrated the initial 15 s was at least as discriminating as the entire video. We evaluated the GP-MCHAT-R, which combines the first 15 s of manually-coded gaze preference (GP) video with M-CHAT-R results on 73 TD children and 28 children with ASD, 36–99 months of age. The GP-MCHAT-R (AUC = 0.89 (95%CI: 0.82–0.95)), performed significantly better than the MCHAT-R (AUC = 0.78 (95%CI: 0.71–0.85)) and gaze preference (AUC = 0.76 (95%CI: 0.64–0.88)) alone. This tool may enable early screening for ASD in RLS.

Saccade Latency Provides Evidence for Reduced Face Inversion Effects With Higher Autism Traits

Individuals on the autism spectrum are reported to show impairments in the processing of social information, including aspects of eye-movements towards faces. Abnormalities in basic-level visual processing are also reported. In the current study, we sought to determine if the latency of saccades made towards social targets (faces) in a natural scene as opposed to inanimate targets (cars) would be related to sub-clinical autism traits (ATs) in individuals drawn from a neurotypical population. The effect of stimulus inversion was also examined given that difficulties with processing inverted faces are thought to be a function of face expertise. No group differences in saccadic latency were established for face or car targets, regardless of image orientation. However, as expected, we found that individuals with higher autism-like traits did not demonstrate a saccadic face inversion effect, but those with lower autism-like traits did. Neither group showed a car inversion effect. Thus, these results suggest that neurotypical individuals with high autism-like traits also show anomalies in detecting and orienting to faces. In particular, the reduced saccadic face inversion effect established in these participants with high ATs suggests that speed of visual processing and orienting towards faces may be associated with the social difficulties found across the broader autism spectrum.

Ophthalmological findings in children with autism spectrum disorder

PURPOSE

Eye pathology could be related to atypical visual behaviours and impaired social communication through visual cues in children with autism spectrum disorder (ASD). The main purpose of this prospective study was to assess ophthalmological disorders in children with ASD and to investigate the relationships with intellectual disability (ID) and ASD severity.

METHODS

In this prospective study, comprehensive ophthalmological and oculomotor examinations were performed. ASD severity and verbal and performance intelligence quotients were determined using adapted scales. These clinical data were compared between groups of children based on the presence or absence of ophthalmological disorders and the achievement or not of visual acuity (VA) testing by using non-parametric statistical tests.

RESULTS

Amongst a sample of 51 children, ophthalmological disorders were found in 39% of cases, with 35% having significant refractive errors and 10% presenting with strabismus. Children with ASD and ophthalmological disorders had significantly lower verbal (29.8 ± 14.7 compared with 44.3 ± 21.5; p = 0.010) and performance quotients (57.8 ± 18.3 compared with 67.59 ± 20; p = 0.049) but no significant result was found between the presence of ophthalmological disorders and ASD severity, level of communication and social contact, or modulating behaviour when changes occur. Children who did not achieve monocular VA testing (39%) had significantly lower verbal (25.1 ± 9.7 compared with 46.1 ± 20.9; p < 0.001) and performance quotients (52.7 ± 17 compared with 69.8 ± 18.8; p = 0.001), also presented higher social interaction impairment (p = 0.002), and expressed more important behavioural signs (p = 0.007).

CONCLUSIONS

Ophthalmological disorders are frequently found in children with ASD, especially in those with ID. Ophthalmologists and child psychiatrists should pay attention to perform ophthalmological examination in children with ASD since eye disorders might remain undetected. A comprehensive examination by a paediatric ophthalmologist would help to improve the individual clinical description and the global intervention.

TRIAL REGISTRATION

Clinical trial registration number: NCT02444117.

Inflexibility in Autism Spectrum Disorder: Need for certainty and atypical emotion processing share the blame

Although ASD (Autism Spectrum Disorder) diagnosis requires the co-occurrence of socio-emotional deficits and inflexible behaviors, the interaction between these two domains remains unexplored. We used an emotional Wisconsin Card Sorting Test adapted to fMRI to explore this question. ASD and control participants matched a central card (a face) with one of four surrounding cards according to one of three rules: frame color, facial identity or expression. Feedback informed participants on whether to change or maintain the current sorting rule. For each rule, we modeled feedback onsets to change, switch (confirming the newly found rule) and maintenance events. "Bias error", which measures participants' willingness to switch, was larger in ASD participants for the emotional sorting rule. Brain activity to change events showed no group differences. In response to switch events significantly larger activity was observed for ASD participants in bilateral Inferior Parietal Sulci. Inflexibility in ASD appears characterized by the unwillingness to switch toward processing socio-emotional information, rather than a major disruption in cognitive flexibility. However, a larger activity to switch events in ASD highlights the need for a higher level of certainty before setting into a stable processing stage, which may be particularly detrimental in the highly changeable socio-emotional environment.