This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on https://www.researchprotocols.org, as well as this copyright and license information must be included.
Impaired facial emotion expression recognition (FEER) has typically been considered a correlate of autism spectrum disorder (ASD). Now, the alexithymia hypothesis is suggesting that this emotion processing problem is instead related to alexithymia, which frequently co-occurs with ASD. By combining predictive coding theories of ASD and simulation theories of emotion recognition, it is suggested that facial mimicry may improve the training of FEER in ASD and alexithymia.
This study aims to evaluate a novel mimicry task to improve FEER in adults with and without ASD and alexithymia. Additionally, this study will aim to determine the contributions of alexithymia and ASD to FEER ability and assess which of these 2 populations benefit from this training task.
Recruitment will primarily take place through an ASD community group with emphasis put on snowball recruiting. Included will be 64 consenting adults equally divided between participants without an ASD and participants with an ASD. Participants will be screened online using the Kessler Psychological Distress Scale (K-10; cut-off score of 22), Autism Spectrum Quotient (AQ-10), and Toronto Alexithymia Scale (TAS-20) followed by a clinical interview with a provisional psychologist at the Federation University psychology clinic. The clinical interview will include assessment of ability, anxiety, and depression as well as discussion of past ASD diagnosis and confirmatory administration of the Autism Mental Status Exam (AMSE). Following the clinical interview, the participant will complete the Bermond-Vorst Alexithymia Questionnaire (BVAQ) and then undertake a baseline assessment of FEER. Consenting participants will then be assigned using a permuted blocked randomization method into either the control task condition or the mimicry task condition. A brief measure of satisfaction of the task and a debriefing session will conclude the study.
The study has Federation University Human Research Ethics Committee approval and is registered with the Australian New Zealand Clinical Trials. Participant recruitment is predicted to begin in the third quarter of 2021.
This study will be the first to evaluate the use of a novel facial mimicry task condition to increase FEER in adults with ASD and alexithymia. If efficacious, this task could prove useful as a cost-effective adjunct intervention that could be used at home and thus remove barriers to entry. This study will also explore the unique effectiveness of this task in people without an ASD, with an ASD, and with alexithymia.
Australian New Zealand Clinical Trial Registry ACTRN12619000705189p; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=377455
PRR1-10.2196/24543
Behavioral and cognitive neuroscience has elaborated a range of emotion-related difficulties with emotion recognition [
The 2 forms of alexithymia, type 1 and type 2 [
Simulation theories [
Alexithymia can explain the failure of this simulation process. Studies suggest that reduced emotion recognition accuracy in alexithymia is related to reduced activity in the insula (a small region within the cerebral cortex) [
The simulation process may also fail in alexithymia due to an inability to access the emotional information generated by a simulation. There are 2 explanations as to why people with alexithymia may not be able to access the simulation information. The first explanation is because of the inability to describe emotions in the self and differentiate them from bodily sensations, which is the hallmark of alexithymia [
Another way that alexithymia may explain a failure of the simulation process is if people with alexithymia have trouble assigning the simulated emotion as belonging to the “sender” of the emotional expression [
While alexithymia is a key element in explaining the failure of this simulation process, it does not provide a finer mechanistic explanation. However, theories proposed within predictive coding frameworks (PCFs) are beginning to provide a compelling explanation.
PCFs suggest that the brain uses complex mathematical models to predict the most likely truths occurring in ourselves and our worlds [
Therefore, when generating a model of the current state of the self or body, oxytocin mediates how much weight will be given to an interoceptive stimulus versus predictive models of the body’s current state. Second, oxytocin has a role in enabling the plasticity required for the brain to generate a model of the emotional and social “self.” Without these models, an over reliance would be put on bottom-up information, as described by Pellicano and Burr [
Two constructs implicated in forming a distinction between the self and others were introduced in the earlier paragraphs: interoception and the self. Interoception is the perception of one’s own internal state, including the perception of sensations such as thirst, stress, temperature, sleepiness, and heartbeat. One use of these sensations is for inferring emotional states and generating a model of the embodied “self” [
Facial mimicry and oxytocin may aid these interoceptive and emotional processing symptoms. Aoki and colleagues [
One natural way to increase oxytocin has been suggested by Delaveau et al [
Lewis and Dunn [
The first aim of this study is to evaluate the efficacy of an instructed mimicry task condition to improve FEER compared to a control condition and also from baseline to posttest in adults with and without alexithymia and ASD. The second aim of the study is to confirm findings by Cook et al [
The study design is a mixed factor 2x2x2 within-between design. Participants who meet inclusion and exclusion criteria will be randomized to a control task or the experimental task. Inclusion and exclusion criteria and pretrial control variables will be collected during an online screening stage and during an in-person clinical interview. Baseline outcomes will be assessed directly before the control task or experimental task. They will then be assessed again during the control and experimental tasks. The study will conclude with a brief measure of satisfaction of the task and a debriefing session. See
We will recruit through local ASD support groups, online social media, and noticeboard advertisements (eg, at Federation University Australia). We will also use snowball recruiting to better penetrate the ASD community. Participants with an ASD (based on the assessments described in the Assessment section) will be assigned to the ASD group. Other participants will be assigned to the alexithymia-matched general population group (AM group).
The researchers will also use the 20-item Toronto Alexithymia Scale (TAS-20) total scores [
The inclusion criterion is that participants are 18 years of age and older. The ASD group also has the added inclusion criterion of a confirmed ASD based on the assessment described in the Assessment section. Exclusion criteria include the participant having another mental health condition (other than an ASD if in the ASD group) diagnosed by a clinician and that the Kessler Psychological Distress Scale (K10) [
The main aim of the current study is to assess the efficacy of the intervention in the general population and an ASD population, but also if people with high levels of alexithymia gain additional benefit. Further, this study will aim to replicate the study by Cook et al [
Participants will then be invited to a clinical interview with a provisional psychologist (PP) at the Federation University Psychology Clinic who will confirm a reported diagnosis of ASD and complete an observational assessment that gives a diagnostic indication of an ASD. Participants who have reported a diagnosis will be asked what instrument was used and if the PP can request a copy of the results. The PP will also complete the Autism Mental Status Exam 2.2 (AMSE) [
The PPs will have pre-existing training in conducting clinical interviews, including mental status examination, as well as administration of symptom measures and cognitive testing. They will be supervised by approved supervisors and endorsed clinical psychologists with extensive clinical experience (at least 10 years) working with and diagnosing people who have an ASD. All psychologists will participate in the AMSE training [
Clinical interviews will be recorded, and AMSE scoring decisions will be made with one or, where it is difficult to reach consensus, both of the supervisors. Participants who meet AMSE criteria and have a confirmed historical diagnosis will be included in the ASD group.
During the screening stage of the study, participants complete the K-10, TAS-20, AQ-10, and several single-item questions pertaining to demographic variables and known presence of other mental conditions. In the clinical interview with the PP, participants will be administered the DASS-21 and WASI-II, and the clinician will also use this time to complete the AMSE observational measure. Indication of context and time-point for each measure can be seen in
Study design and measures.
Concept | Measure | Instrument | Condition/ time point | ||||
Online screening stage | Clinical interview stage | Experiment stage | |||||
|
|
|
|
|
Pretest | During Experimental task | Posttest |
Primary outcome | FEERa ability | FEER task Improvement score | - | - | X | X | - |
Secondary Outcome | Acceptance / Enjoyment of task | Task satisfaction Questionnaire | - | - | - | - | X |
Eligibility | Psychological distress | K-10b | X | - | - | - | - |
Depression and Anxiety | DASS-21c | - | X | - | - | - | |
Predict ASDd classification on the ADOSe | AMSEf | - | X | - | - | - | |
Predict ASD classification | AQ-10g | X | - | - | - | - | |
Alexithymia | TAS-20h | X | - | - | - | - | |
Control variables | Alexithymia | BVAQi | - | - | X | - | - |
IQ | WASI-2j | - | X | - | - | - | |
Demographics | Demographic information | X | - | - | - | - |
aFEER: facial emotion expression recognition.
bK-10: Kessler Psychological Distress Scale.
cDASS-21: Depression Anxiety Stress Scales-21.
dASD: autism spectrum disorder.
eADOS: Autism Diagnostic Observation Schedule.
fAMSE: Autism Mental Status Exam.
gAQ-10: Autism Spectrum Quotient.
hTAS: Toronto Alexithymia Scale.
iBVAQ: Bermond–Vorst Alexithymia Questionnaire.
jWASI-2: Wechsler Abbreviated Scale of Intelligence.
The K-10 [
The 20-item TAS-20 [
The Bermond–Vorst Alexithymia Questionnaire (BVAQ) [
The DASS-21 [
The AMSE 2.2 [
The WASI-II [
The AQ-10 [
In addition to these measures, a self-developed questionnaire of satisfaction, acceptance, and enjoyment of the task will also be designed to assess how viable this method of training is for people with alexithymia and ASD (
The Tobii Pro x3-120 Eye Tracker (TPX3) is a screen-based eye tracker that tracks the gaze point on a screen. The TPX3 has been certified to meet European standard EN 62471, which indicates that it is not harmful to the human eye. The TPX3 has a sampling rate of 120 Hz and a large head movement box (area where the user’s head can move and still be tracked). Only the gaze point 2D coordinates on the screen for each eye will be used. The TPX3 will attach to the bottom of a computer monitor that is ~64 cm in front of the participant. Both the chair and the participant’s screen can adjust for height to make sure the participant’s head is within the head movement box as indicated in
Physical layout of experiment and eye-tracker.
The primary outcome being assessed is an increase in FEER. This will be measured by comparing the percent of correctly identified facial emotion expressions from the baseline assessment task to the instructed mimicry experimental task. Additionally, the percent of correctly identified expressions will also be compared between the control condition and the intervention condition. This will provide 2 primary metrics: before-after percent improvement and control group to experimental group percent improvement.
The task will be delivered in-person at the Federation University Psychology Clinic in Ballarat, Australia, by the primary author in a small counselling room. The room will have 2 computers on a desk, each with separate monitors and peripherals (one for the participant and the other for recording eye-tracking data). Participants will take part in the task 1 participant at a time based on a mutually agreed upon time and date. Participants will only undergo this task once, and it is expected to take approximately 30 minutes based on prior testing where a small neurotypical group took approximately 25 minutes to complete the task in a usability test. The intervention is a software-based FEER training task that presents a series of different facial expressions and requires individuals to firstly mimic and then select which emotional expression they saw. The software runs on Microsoft Windows 10 and requires only minimal system requirements and a mouse.
The intervention task condition will begin with the software presenting a basic description of 6 facial expressions (fear, sadness, happiness, anger, disgust, surprise) and tips on how to spot each expression. Next, participants will be verbally advised by the experimenter and by on-screen text that a series of emotion recognition trials will begin shortly. Further, they will be told that after each emotion is displayed that they should try their best to mimic (copy) the observed facial expression with their own face. Once they have mimicked the expression, they should then use the 6 labelled buttons on the right-hand side (or associated keyboard shortcut) to choose which emotion was displayed. After selecting the answer using 1 of the 6 buttons, a 5-point Likert scale on the computer screen will ask the participant “How well did you mimic the expression?” with responses ranging from 1 “not very well” to 5 “very well.” This will repeat for each of the 36 trials. The trials will include various models of both men and women depicting the 6 emotions.
The sequence of each trial in the task is depicted in
In a task that aims to test or train FEER ability, it is important to ensure that the emotional stimuli tested against are true exemplars of the emotion they purport to signal. It is also important that the expressive images in the baseline test and the experimental tasks are similar in recognition difficulty. This will reduce the risk of improvements being attributed to easier expressions in the experimental tasks compared to the baseline task. This task used the Warsaw Set of Emotional Facial Expression Pictures (WSEFEP) [
Task sequence and stimulus example.
This study includes both a before-after, within-group comparator and a between-groups control group. To make the before-after comparison, participants will first partake in a baseline assessment task of emotion recognition before randomization to one of the experimental conditions. This baseline assessment task will be the same as the intervention task; however, participants will not be instructed to mimic the facial expressions prior to selecting an expression.
In addition to the before-after comparator, there will be a control condition. The control condition will also undergo the baseline assessment task; however, instead of participating in the mimicry intervention task, they will undergo a similar task with no mimicry element. This task will be identical to the instructed mimicry intervention task; however, instead of being instructed to mimic the facial expression, they will be instructed to “pay particular attention to the entire face prior to selecting an expression.” This instruction is aimed to match for attention levels.
A final point of comparison will be that participants will be first broken up into 2 groups, namely the AM group and ASD group. This comparison will provide the ability to analyze if any intervention effects only occur or are stronger in people with ASD, alexithymia, or the general population. The participant flow diagram can be seen in
Participant flow diagram. ASD: autism spectrum disorder; K-10: Kessler Psychological Distress Scale.
Participants will be randomized using permuted block randomization with stratification across age and gender to receive either the control or experimental task condition with equal numbers from both the ASD group and AM group receiving each task.
Based on the effect sizes of 2 previous studies measuring similar constructs and similar tasks [
The data will be analyzed using SPSS, Version 25. Alexithymia will be analyzed to aid in matching the AM group and ASD group on alexithymia. To ensure that the ASD and AM groups, and also the experimental and control groups, are comparable, means and standard deviations of participant characteristics will also be compared with tests of difference.
The study design is a mixed factor, 2x2x2, within-between design (see
In addition to the mixed factorial ANOVA, a hierarchical regression analysis will also be conducted (see
Experimental group and intervention design.
Time point | ASDa group (n=~32) | AMb group (n=~32) | ||
|
Control task condition (n=~16) | Mimicry task condition (n=~16) | Control task condition (n=~16) | Mimicry task condition (n=~16) |
Pretest | FEERc task score | FEER task score | FEER task score | FEER task score |
Posttest (the intervention task is the posttest) | FEER task score | FEER task score | FEER task score | FEER task score |
aASD: autism spectrum disorder.
bAM: alexithymia-matched general population.
cFEER: facial emotion expression recognition.
Hierarchal regression stages for analysis.
Stage | Variables |
Stage 1 | IQ, age, gender |
Stage 2 | +Autism |
Stage 3 | +Alexithymia |
Stage 4 | +Intervention |
It is hypothesized that the condition that receives the mimicry task will show significantly better emotion recognition improvement scores (posttest minus baseline); the degree of alexithymia will negatively predict baseline scores; the degree of alexithymia will predict improvement scores (posttest minus baseline); when matched on alexithymia, the general population and ASD population will not significantly differ in emotion recognition improvement scores (posttest minus baseline); and when matched on alexithymia, the AM group and ASD group will not significantly differ on baseline emotion recognition scores.
This study will meet the ethical guidelines outlined in the National Statement on Ethical Conduct in Human Research (NHMRC, ARC & UA) [
As of June 2021, the project has not begun recruitment. The project is predicted to begin participant recruitment in the third quarter of 2021, with data collection beginning as eligible consenting participants make contact.
This paper describes, to our knowledge, the first study to assess an instructed mimicry task to increase emotion recognition ability in people with alexithymia and ASD. Despite the formulation of the alexithymia hypothesis, very few studies have assessed the impact of alexithymia in interventions aimed at increasing social functioning and in particular, emotion recognition. Additionally, to our knowledge, no other study has explored the benefits of instructed mimicry on emotion recognition in people with ASD and alexithymia. This mixed factor, 2x2x2, within-between design will assess the efficacy of the instructed mimicry task to increase FEER. This study design will enable the assessment of baseline and improvement scores individually for people with and without ASD while controlling for levels of alexithymia. This will allow analyses of how this intervention uniquely benefits ASD without the interference of alexithymia, which is known to frequently co-occur in ASD [
One proposed benefit of this intervention for people with alexithymia and ASD is that the software does not require supervision. This will likely benefit these populations as they typically enjoy computer-based interaction rather than social interaction, as indicated by previous studies (eg, [
Despite the benefits of this task and study design, some limitations are present. First, matching the ASD and AM groups on levels of alexithymia improves our ability to exclude alexithymia as a contributing factor of emotion recognition accuracy. Unfortunately, it also means that other confounding factors may have been systematically introduced when matching on alexithymia. This limitation will make it difficult to attribute any benefits of the intervention task specifically to autism. Additionally, when comparing emotion recognition accuracy, it will limit our ability to suggest whether emotion recognition difficulties are related to ASD or to alexithymia, as per the alexithymia hypothesis.
A second limitation is the once-off nature of this intervention. While similar tasks have shown improvement over one session (eg, [
In summary, it has been suggested that emotion recognition problems in ASD are instead due to co-occurring alexithymia. This study matches an ASD and non-ASD general population group (the AM group) on alexithymia to improve the ability to detect if this is true. Further, this study design will aid in detecting if the instructed mimicry intervention benefits people specifically with ASD or if it more broadly aids people with alexithymia. If successful, this adjunct intervention will be cost-effective and easy to implement while also removing entry barriers as it could be used at home on a personal computer.
Workflow of participants through in-person segments of study.
Questionnaire of Task Satisfaction.
Selection process of pictures to match task difficulty between the baseline task and the experimental task.
Autism Diagnostic Observation Schedule
alexithymia-matched general population
Autism Mental Status Exam
analysis of variance
Autism Spectrum Quotient
autism spectrum disorder
Depression Anxiety Stress Scales-21
facial emotion expression recognition
Kessler Psychological Distress Scale
predictive coding framework
provisional psychologist
Toronto Alexithymia Scale
Wechsler Adult Intelligence Scale - Fourth Edition
Wechsler Abbreviated Scale of Intelligence
Warsaw Set of Emotional Facial Expression Pictures
We thank and acknowledge the Australian Government Research Training Program for supporting the first author.
None declared.