Pilot Study to Identify Candidate Biomarkers for Autism Spectrum Disorder Based on Perception and Production of Facial Expressions

Core Concepts

Facial expression perception and production may serve as potential stratification biomarkers to explain the heterogeneity of autism spectrum disorder.

Abstract

This pilot study aims to identify candidate biomarkers for autism spectrum disorder (ASD) based on the perception and production of facial expressions. The study includes 11 children and young adults diagnosed with ASD and 11 age- and gender-matched neurotypical (NT) individuals. Participants complete recognition and mimicry tasks using customizable 3D avatar stimuli while their webcam captures eye tracking and facial video tracking data.
The key highlights and insights from the study are:

Data Acquisition and Imputation:

Significant data loss is observed, particularly for eye tracking and video tracking in the later stages of the experiment.
The study uses robust K-nearest neighbors imputation to handle missing data.

Construct Validity:

220 out of 312 dependent variables (DVs) have valid constructs based on the expected responses in the NT group.
Constructs are evaluated for measures of expression recognition accuracy, gaze preference to the face, expression mimicry, and facial action unit activation and asymmetry.

Group Discriminability:

Using the Boruta algorithm, one candidate biomarker is identified: the percentage of gaze duration to the face while mimicking a 'disgust' expression in the uncustomized static condition.
An additional 14 DVs are identified as potentially relevant for group discrimination between ASD and NT using the modified r-Boruta algorithm.

Power Analysis:

Power analysis is conducted to provide sample size recommendations for future studies based on the observed effect sizes of the identified candidate biomarkers and DVs of interest.

This pilot study provides a framework for ASD stratification biomarker discovery based on the perception and production of facial expressions. The findings highlight the potential of facial expression processing as a target for identifying subgroups within the autism spectrum.

Stats

The percentage of gaze duration to the face while mimicking a 'disgust' expression in the uncustomized static condition is lower in the ASD group (M=0.41, SD=0.18) compared to the NT group (M=0.63, SD=0.14).
The activation of Action Unit 6 (cheek raiser) while mimicking a 'happy' expression in the uncustomized dynamic condition is lower in the ASD group (M=0.41, SD=0.25) compared to the NT group (M=0.63, SD=0.18).
The asymmetry of Action Unit 12 (lip corner puller) while mimicking a 'happy' expression in the customized dynamic condition is greater in the ASD group (M=0.11, SD=0.09) compared to the NT group (M=0.05, SD=0.04).

Quotes

"Facial expression production and perception in autism spectrum disorder (ASD) suggest potential presence of behavioral biomarkers that may stratify individuals on the spectrum into prognostic or treatment subgroups."
"Construct validity and group discriminability have been recommended as criteria for identification of candidate stratification biomarkers."

Key Insights Distilled From

Pilot Study to Discover Candidate Biomarkers for Autism based on Perception and Production of Facial Expressions

by Megan A. Wit... at arxiv.org 04-26-2024

https://arxiv.org/pdf/2404.16040.pdf

Deeper Inquiries

How can the findings from this pilot study be leveraged to develop more ecologically valid and engaging facial expression tasks for individuals with ASD?

The findings from this pilot study provide valuable insights into the perception and production of facial expressions in individuals with ASD. To develop more ecologically valid and engaging facial expression tasks for individuals with ASD, the following strategies can be considered:

Customization of Avatar Stimuli: The use of customizable 3D avatars in the study proved to be engaging for participants. Further customization options, such as allowing participants to personalize their avatars or choose from a variety of avatars, can enhance engagement and motivation.

Dynamic and Interactive Tasks: Incorporating dynamic and interactive elements into facial expression tasks can make them more engaging for individuals with ASD. For example, using avatars that respond to the participant's expressions in real-time or incorporating gamified elements can increase engagement.

Incorporating Real-Life Scenarios: Designing tasks that simulate real-life social interactions or scenarios can enhance the ecological validity of facial expression tasks. For example, creating tasks that involve interpreting facial expressions in social contexts like conversations or group settings can make the tasks more relevant and engaging.

Multi-Modal Approach: Combining facial expression tasks with other modalities, such as voice tone analysis or body language recognition, can provide a more comprehensive understanding of social communication skills in individuals with ASD. This multi-modal approach can make the tasks more engaging and reflective of real-world social interactions.

Feedback and Reinforcement: Providing immediate feedback and positive reinforcement during the tasks can enhance engagement and motivation. Visual or auditory cues indicating correct responses or progress can make the tasks more interactive and rewarding for participants.

By incorporating these strategies based on the findings of the pilot study, researchers can develop more ecologically valid and engaging facial expression tasks for individuals with ASD, ultimately improving the effectiveness of biomarker discovery and stratification in ASD research.

What other behavioral or physiological measures, in addition to facial expression processing, could be explored as potential stratification biomarkers for ASD?

In addition to facial expression processing, several other behavioral and physiological measures can be explored as potential stratification biomarkers for ASD. Some of these measures include:

Eye Tracking Patterns: Eye tracking data can provide insights into visual attention and social cognition in individuals with ASD. Patterns of gaze fixation, saccades, and pupil dilation can serve as biomarkers for social communication difficulties and sensory processing differences in ASD.

Heart Rate Variability (HRV): HRV, which reflects the variation in time intervals between heartbeats, can be indicative of autonomic nervous system function and emotional regulation. Altered HRV patterns have been observed in individuals with ASD and can serve as a biomarker for emotional dysregulation and stress responses.

Electrodermal Activity (EDA): EDA, also known as skin conductance, measures changes in sweat gland activity and is linked to emotional arousal and stress. Abnormal EDA responses have been reported in individuals with ASD and can be explored as a biomarker for emotional reactivity and regulation.

Neuroimaging Markers: Structural and functional neuroimaging techniques, such as MRI and fMRI, can reveal differences in brain structure and connectivity associated with ASD. Biomarkers derived from neuroimaging data, such as amygdala activation patterns or white matter integrity, can provide insights into the neural correlates of ASD.

Behavioral Response to Social Stimuli: Observing and analyzing behavioral responses to social stimuli, such as social interactions, facial expressions, or emotional cues, can offer valuable information about social communication deficits in individuals with ASD. Quantifying social behaviors and responses can serve as biomarkers for social cognition and interaction difficulties.

By exploring a combination of these behavioral and physiological measures alongside facial expression processing, researchers can identify a comprehensive set of potential stratification biomarkers for ASD, enhancing the understanding and characterization of the disorder.

How might the identified candidate biomarkers and DVs of interest relate to the underlying neurobiology and cognitive processes associated with ASD?

The identified candidate biomarkers and DVs of interest, such as percentage gaze duration to the face while mimicking 'disgust' expression and activation of specific action units during facial expression mimicry, can provide valuable insights into the underlying neurobiology and cognitive processes associated with ASD.

Neurobiological Correlates: The candidate biomarkers related to facial expression processing may reflect differences in neural processing and connectivity in individuals with ASD. For example, altered activation patterns in brain regions involved in emotion recognition and social cognition, such as the amygdala and fusiform gyrus, may underlie difficulties in facial expression perception and production in ASD.

Social Cognitive Processes: The DVs of interest, including mimicry accuracy and asymmetry of action units, can shed light on social cognitive processes in individuals with ASD. Differences in mimicry abilities and facial expression production may be linked to challenges in social communication, theory of mind, and emotional regulation characteristic of ASD.

Emotional Regulation: Biomarkers related to gaze patterns, heart rate variability, and electrodermal activity can provide insights into emotional regulation and arousal in individuals with ASD. Dysregulated emotional responses and atypical physiological reactions to social stimuli may be associated with the emotional processing difficulties observed in ASD.

Sensory Processing: Behavioral responses to social stimuli and eye tracking patterns can be indicative of sensory processing differences in individuals with ASD. Hypersensitivity or hyposensitivity to social cues and facial expressions may be reflected in the identified biomarkers and DVs, highlighting the role of sensory processing in ASD.

By examining the relationship between the identified biomarkers and DVs and the underlying neurobiology and cognitive processes associated with ASD, researchers can gain a deeper understanding of the mechanisms driving social communication deficits and inform targeted interventions and treatments for individuals with ASD.