Published: October 4th, 2018
This protocol is designed to measure reward anticipation and processing in young children with and without autism. Specifically, the protocol is designed to study the neural correlates of reward during social and nonsocial conditions while controlling for reward between conditions.
We present a protocol designed to measure the neural correlates of reward in children. The protocol allows researchers to measure both reward anticipation and processing. Its purpose is to create a reward task that is appropriate for young children with and without autism while controlling reward properties between two conditions: social and nonsocial. The current protocol allows for comparisons of brain activity between social and nonsocial reward conditions while keeping the reward itself identical between conditions. Using this protocol, we found evidence that neurotypical children demonstrate enhanced anticipatory brain activity during the social condition. Furthermore, we found that neurotypical children anticipate social reward more robustly than children with autism diagnoses. As the task uses snacks as a reward, it is most appropriate for young children. However, the protocol may be adapted for use with adolescent or adult populations if snacks are replaced by monetary incentives. The protocol is designed to measure electrophysiological events (event-related potentials), but it may be customized for use with eye-tracking or fMRI.
Autism spectrum disorder (ASD) is a developmental disability characterized by impairments in social communication (verbal and non-verbal) and the presence of restricted interest and/or repetitive behaviors1. Given that ASD is hypothesized to be neurologically-based2,3, it is unsurprising that neuroscience research involving children with ASD has become highly prevalent over the past decade. Though many theories about the brain basis of ASD have been proposed, one in particular that has garnered considerable research attention is the social motivation hypothesis4. Briefly, the social motivation hypothesis states that children with ASD engage in less social interaction than their typically developing (TD) peers because social interaction is not as rewarding for them. Chevallier et al. provide a review of the social motivation hypothesis5. Because this hypothesis directly relates to the reward system, specifically whether or not the system in ASD is responsive to social information, multiple studies have investigated the social reward system in ASD6,7,8,9,10,11,12. Results from these studies have differed, with some providing evidence that the reward system in ASD is hypoactive to both social and nonsocial information, and others suggesting that the reward system functions typically for nonsocial information but is hypoactive to social stimuli. One potential reason for these inconsistent results relates to the stimuli and methodology used in the protocols. It is difficult to match social and nonsocial rewards in an experimental context; for example, multiple studies have used a picture of a smiling face as the social reward, and the nonsocial reward is monetary (e.g., getting money after the experiment is complete7,8,11). Though these studies provide an important foundation for future research, it is difficult to determine whether or not the findings relate to differences in social versus nonsocial reward responsiveness in ASD or if they are due to differences between reward conditions.
The current protocol is designed to investigate the reward system in high-functioning children with ASD using electrophysiology. To explore differences between children with and without ASD based on reward anticipation, the stimulus-preceding negativity (SPN) was measured. The SPN is a slow-wave component that reflects an expectation of a reward stimulus13. The significance of the SPN is typically conceptualized as emotional anticipation14,15,16 and is thought to be reflected by activity in the insula17,18. The SPN is often measured after participants perform a motor response and before feedback onset during a decision-making task19,20. The SPN is sensitive to reward magnitude and is consistently larger in reward versus no-reward conditions15,16,21. Though the SPN is typically measured during decision-making tasks, researchers have reported that the SPN can be observed when anticipating affective upcoming stimuli without any task22,23,24. One critical aim of the current protocol is to perform an experimental task in which the rewards between social and nonsocial conditions are matched to eliminate potential confounds. Another goal is to test young children between 6 and 11 years old. Therefore, the protocol can serve as an age-appropriate reward task that children may find engaging without becoming frustrated.
Procedures involving human participants were approved by the Human Subject Research Ethics Committee/Institutional Review Board at University of California, Riverside and University of California, San Diego.
1. Stimuli Preparation
Note: The procedures described below are specific to a commercially available editing suite (see Table of Materials); however, other photo editing software can certainly be used.
2. Behavioral Procedures
3. EEG Recording
4. Processing EEG Data
Note: The procedures and commands described in this section are specific to EEGLAB and ERPlab toolboxes28.
5. Processing Differences for ERSP Analysis
6. Statistical Analysis
Designing experiments to systematically compare brain activity with social versus non-social reward stimuli is complex, due to the inherent difficulty in equating social and non-social rewards. Figure 1 represents stimuli from an experimental protocol designed for investigating neural responses to reward while controlling for reward properties. Specifically, this paradigm was designed to (i) keep rewards consistent between social and nonsocial trials, (ii) control for physical stimulus properties between social and nonsocial trials, and (iii) be age-appropriate for 6 to 11-year-old children with and without autism.
Figure 2 depicts ERP responses as participants anticipate social and non-social stimuli. It should be noted that because the current protocol was designed to measure reward anticipation (the SPN), the displayed epochs are largely prior to feedback onset (which occur at 0 ms in the figures). These results suggest that typically developing (TD) children anticipate reward stimuli accompanied by faces more robustly than children with ASD. Furthermore, though TD children anticipate face stimuli significantly more than non-face stimuli, children with ASD do not evidently show significant differences in brain activity between conditions.
Figure 1: Schematic of stimulus presentation and timing. Feedback for the social (face) condition is shown in the left column. Feedback for the nonsocial (non-face) condition is shown in the right column. Feedback for "correct" answers is shown on top, and feedback for "incorrect" answers is shown below. This figure is re-printed with permission12. Please click here to view a larger version of this figure.
Figure 2: Grand averaged waveforms for TD children and those with ASD from the SPN in response to social/faces (left) and nonsocial/arrows (right). TD children are represented by a solid line and children with ASD by a dashed line. The area between -210 and -10 ms, used for statistical analysis, is highlighted with a gray box. This figure is modified from a previous publication6. Please click here to view a larger version of this figure.
The current article describes the stimuli, data collection process, and analysis of ERP data in a reward paradigm for children. In this paradigm, children play a guessing game similar to pick-a-hand on the computer and see feedback about whether their guess is correct or incorrect. ERP results for reward anticipation (brain activity prior to the onset of feedback) were consistent with the stimulus-preceding negativity (SPN). Between conditions, the results suggest that TD children anticipate reward stimuli accompanied by faces more strongly than reward stimuli accompanied by non-face images6. Between groups of children, the results suggest that TD children anticipate face stimuli significantly more than children with autism do. These results are exciting, as they provide important information about how social and nonsocial information are anticipated in children with autism. This is particularly important for furthering the understanding of neural mechanisms of autism and providing support for the social motivation hypothesis. These findings provide useful information for the creation and refinement of interventions, as it underscores the importance of social motivation for children with ASD; for example, it may be important for interventions to explicitly attempt to increase the reward value of social partners to directly affect social motivation in this population.
This protocol is useful for measuring anticipatory brain activity in children with and without ASD, and the data provides evidence that this type of brain activity can be reliably and successfully elicited in children over 6 years old. Furthermore, this method allows social and nonsocial conditions to be directly compared without the presence of confounds related to reward properties (since the reward for correct responses was goldfish in both conditions). In the current protocol, faces were scrambled and an arrow shape was created. This procedure preserves the physical stimulus properties of faces in the nonsocial (non-face) condition. This protocol may be useful for future investigations into sub-groups of ASD (e.g., some children with ASD are more socially motivated than others), and could be utilized to better understand why some children respond more effectively than others to certain interventions.
There are limitations to the current approach that must be taken into consideration. First, the paradigm described above is useful for children between 6 and 11 years old with and without ASD who have cognitive abilities in the average range. Pilot data of typically developing children younger than 6 was not successful, as children were confused by the directions and did not understand the game's instructions. In the current protocol, exclusionary criteria included a full-scale IQ score below 70. Therefore, the current paradigm may not be appropriate for children with a mental or chronological age below 6. However, it may be possible to modify the current protocol so it is appropriate for individuals with lower IQs and younger children. Some modifications to make it more appropriate for young children such as toddlers are currently being investigated. Such modifications include changing the task to be passive (e.g., having children watch stimuli that appear at predicable intervals in a block design) and using an S1/S2 paradigm24. In such a design, the content of S1 reliably provides information about the content of S2 (e.g., if S1 is a square, then S2 will be a face; if S1 is a circle, then S2 will be an arrow). Alternatively, the timing structure of the current paradigm could be used to create an anticipatory auditory protocol.
In ASD, it would be of interest to use speech versus non-speech groups and measure brain activity in children with ASD who are non-verbal and have difficulty responding to instructions or attending to visual stimuli31. Related to the first limitation, it should be noted that results from children with ASD who have cognitive abilities in the average range are likely not representative of the entire autism spectrum - which, by definition, captures a broad range of functioning levels. Therefore, these representative results cannot be extrapolated to all children with ASD. Finally, it is important to note that the stimuli used in the current protocol were normed by adults rather than children. Therefore, future studies should consider using a stimulus set of facial expressions normed by children.
Open Access fees for this article were provided by Brain Vision LLC.
We thank all the children and families who participated in the protocols described. Publishing fees were paid by Brain Products.
|Electro-Cap - Small (50-54 cm)/Electro-Cap -Small/Extra Small
|EEG Recording Software
|Stimulus Presentation Software
|JMP Pro 11
|Statistical analysis software
|NimStim Face Stimulus Set
|N/A, open source images
|Open source, Available at https://www.macbrain.org/resources.htm
|N/A, free software
|N/A, free software
|EEG analysis software (free download)
|N/A, free software
|N/A, free software
|EEG analysis software (free download)
|Adobe Photoshop, image editing software
|Photoshop 'scramble' plug-in
|photoshop plug-in to scramble images
|NUAMPS EEG AMPLIFIERd
Copyright © 2024 MyJoVE Corporation. All rights reserved