The Specificity of Inhibitory Impairments in Autism and Their Relation to ADHD-type Symptoms
Dublin Core
Title
The Specificity of Inhibitory Impairments in Autism and Their Relation to ADHD-type Symptoms
Creator
Charlotte Sanderson
Date
2010
Description
Findings on inhibitory control in autism have been inconsistent. It is proposed that this may be partly task-related, with different ‘inhibition’ tasks tapping different classes of inhibitory ability. Thus, children with autism (CWA) (N = 31) and typically developing controls (TDC) (N = 28) matched for verbal and non-verbal mental age completed three tasks of inhibitory control, each representing different inhibition subcomponents: a Go/No-Go task (delay inhibition), the Dog-Pig Stroop task (conflict inhibition), and a Flanker task (resistance to distractor inhibition). Behavioural ratings of inattention and hyperactivity/impulsivity were also obtained for each child to consider a possible source of heterogeneity in inhibitory ability. It was predicted that the conflict task would be more problematic for CWA, and that higher ADHD-symptom ratings would predict poorer performance. On the Go/No-Go task, CWA showed superior inhibitory function to controls – making fewer false alarm errors and better task sensitivity. On the Dog-Pig Stroop, CWA showed impaired performance compared to controls – making more accuracy and speed related inhibitory errors. On the Flanker task, CWA showed equivalent inhibitory performance to TD children. Inhibitory impairments were predicted by high ratings of inattention in CWA, but only on the Dog-Pig Stroop. It is argued that CWA are perhaps impaired on tasks of conflict, but not delay or resistance to distractor inhibition. This may reflect the additional working memory demands of these tasks, and suggests that inhibitory difficulty is not a core executive deficit in autism. Symptoms of inattention may be an important predictor of inhibitory heterogeneity amongst CWA.
Subject
inhibition
Stroop
autism
Stroop
autism
Source
Sessions were completed in a well-lit and quiet room, free of distractions. Participants were tested individually, and completed the three inhibitory control tasks (Go/No-Go task; Dog-Pig Stroop task; Flanker task) and two standardised measures (RCPM; BPVS) in counterbalanced order. Experimental session lasted approximately 40-60 minutes.
All three inhibition tasks were written using Psyscript, and run on a computer using an OS X 10.6 operating system.
Go/No-Go Task
Task Design.On each trial, a shape (O, ∆, ⧠, or ◊) would appear centrally on the computer screen. The shapes were simple black line-drawings, subtending approximately 5° vertically and horizontally. Prior to the task, children were instructed to respond to three of the shapes by pressing a large external “star” button (i.e. “Go” stimuli), but to resist responding to a fourth shape (i.e. the “No-Go” stimulus). The shape designated as the “No-Go” stimulus was counterbalanced between participants. To generate a prepotent response, 75% of trials were “Go” trials requiring a button press, and 25% of trials were “No-Go” trials where the response should be withheld.
The maximum inter-stimulus interval (ISI) (i.e. from stimulus onset to stimulus onset) was 2500ms. At the start of each trial, a fixation cross would appear at the centre of the screen for 200ms. This was then replaced by the stimulus, which remained on-screen for 200ms. After the stimulus offset, participants had a further 1000ms to respond, at which point the trial automatically terminated. Stimulus presentation was followed by a 1100ms pause before the next trial commenced. An error tone (“bleep”) was played immediately if the child made an omission error (i.e. failed to respond on a “Go” trial), or a false alarm (i.e. pressed the star button on a “No-Go” trial). A positive feedback-noise (“ping”) was played if the participant made a correct response.
Procedure. Before starting the task, each child completed a warm up session to familiarize with the “Go” and “No-Go” stimuli. Training was terminated only when the child could correctly identify the required response for each shape. Children then completed a short practice block of eight trials containing all four stimuli presented in a fixed, but superficially random order. Then followed 144 experimental trials split into three 48-trial blocks, each separated by a short break. Stimulus presentation was randomised throughout each half block to avoid clustering of “No-Go” trials. The task (including training) lasted approximately ten minutes.
Four measures of task performance were obtained:
1. Number of false alarms (or commission errors): “No-Go” trials on which the button was pressed. This is the main measure of inhibitory control, with false alarms representing failure to inhibit the prepotent button-press response.
2. Number of hits: “Go” trials on which the child responded. This is not a main measure of inhibitory control performance, but indicates how reliably participants detect targets when present and suggest the strength of the prepotent response generated. Hit-rates are also used for calculations of task sensitivity (see below).
3. Task Sensitivity: Estimates of participants’ task sensitivity can be calculated using signal detection theory (A0) and probability estimates of False Alarms and Hits. This permits differentiation between participants who make fewer false alarms, but also fewer hits (poor task sensitivity), and those who make fewer false alarms despite a good hit rate (good task sensitivity). This is important because a low false alarm rate could be due to a generally low response rate (for both “Go” and “No Go” stimuli). Task sensitivity (A0) is a nonparametric measure which ranges from 0.5 (chance performance) to 1 (perfect sensitivity), and is calculated as follows (Grier, 1971):
A = 0.5 (H-FA) (1+H-FA) / [4H (1-FA)]
Where, H = probability (Hits), FA = probability (False Alarms).
4. Hit Trial Reaction Time (RT): Although not a measure of inhibitory control per se, this might indicate between-group differences in processing speed and/or task-strategy.
Dog-Pig Stroop Task.
Task Design. The stimuli used in this task were two simple line drawings of a dog and a pig (see Appendix 4). Stimuli were presented centrally on the computer screen, subtending approximately 6° vertically and 9° horizontally. Two experimental conditions, each containing 32 trials were administered. In the control (baseline) condition, children were simply instructed to say “dog” when they see the dog-image, and “pig” when they see a pig as quickly possible. In the Stroop (i.e. inhibition) condition, children were instructed to say "dog" to pig images, and "pig" to dog-images, as quickly and accurately as possible.
Children’s responses were recorded by an assistant during the task, and also audiotaped so that manuscripts could be subsequently checked by the experimenter. If a child made a mistake on a trial and then corrected themselves, their initial response was recorded. To estimate response latency on each trial, the experimenter would press a large external button as soon as the child made their initial response. Although this measure of reaction time is relatively crude, many of the children taking part would not have been testable with throat microphones which measure voice-onset. These technologies are highly sensitive to all sounds including subtle body movements, lip-smacks and vocalizations, reducing their reliability for use with participants who might have difficulty minimizing task-irrelevant movement or vocalizations. It is also notable that the additional error in reaction-time estimates induced by this method would be constant across groups.
On each trial, the stimulus remained centrally on-screen until a response had been registered (i.e. the response button had been pressed). If no response had been registered after 3000ms had elapsed, the trial automatically terminated, and the message “Too Slow” was presented for 500ms. Stimulus presentation was followed by a 2000ms pause (inter-trial interval) before the next trial commenced. The maximum ISI was thus 5500ms.
Procedure. All children completed the control condition first to provide a measure of baseline picture naming speed and accuracy[2]. After the control condition had been completed, children were presented with training slides to familiarise them with the Stroop naming procedure. After successfully completing the four practice trials, children would commence the 32-trial Stroop condition block. The task (including training) lasted approximately 7 minutes.
Flanker Task
Task Design. For this computer task, children were presented with two large arrow-shaped buttons – one pointing left and one pointing right. There were three experimental conditions: baseline, congruent, and incongruent. Children were asked to respond by pressing the arrow-button pointing the same way as the white target arrow, which was positioned centrally, subtending approximately 4° vertically and 6° horizontally. On baseline trials, the white target arrow was presented on its own. On congruent trials, the white target arrow was flanked by four red ‘distractor’ arrows pointing the same way as the target (e.g. ààààà). On incongruent trials, the white target arrow was flanked by four red ‘distractor’ arrows facing in the opposite direction to the target arrow (e.g. ßßàßß). It is thus only on incongruent trials that the distractors must be actively inhibited/suppressed for correct target identification.
The maximum ISI was 2900ms. A fixation cross would appear centrally on-screen for 200ms. This was then replaced by the stimulus (neutral, congruent or incongruent), which remained on-screen until a button-press had been registered. If no response had been registered after 1200ms had elapsed, the trial automatically terminated. An error-tone (“bleep”) was played if the participant pressed the wrong arrow-button . If the child failed to respond before the trial terminated, an error-tone was played and a “Too-Slow” message was briefly displayed. When the child responded correctly, a positive feedback-noise was given (a “ping”). There was a 1100ms pause (inter-trial interval) between trials.
Procedure. Each child first completed a series of familiarisation trials. This was followed by three blocks of 30 trials separated by a short break (90 trials in total). Each block contained ten baseline, ten congruent and ten distractor trials, which were distributed randomly. Error-rates and mean reaction times (RT) for neutral, congruent and incongruent trials were recorded.
[1] Although a cut-off of 30-points is typically used with younger children, a slightly lower cut-off score is thought to be more accurate for use with older children/adolescents (Mesibov et al., 1989). This is due to the inclusion of one or two items on which older children with autism tend not to score highly (e.g. imitation).
[2]Condition-order was fixed because a pilot study showed that if children completed the experimental (i.e. Stroop) condition first they had difficulty forgetting the ‘opposite’ rule in order to name the pictures normally for the control condition. This was shown by elevated error-rates and poorer naming speeds. Therefore, in order to obtain a realistic measure of ‘automatic’ (i.e. control-condition) picture naming speed and accuracy, and a stronger prepotent response, it was decided appropriate to fix the order of condition presentation (Control, then Stroop). Although this may lead to practice effects, this effect is constant across groups.
All three inhibition tasks were written using Psyscript, and run on a computer using an OS X 10.6 operating system.
Go/No-Go Task
Task Design.On each trial, a shape (O, ∆, ⧠, or ◊) would appear centrally on the computer screen. The shapes were simple black line-drawings, subtending approximately 5° vertically and horizontally. Prior to the task, children were instructed to respond to three of the shapes by pressing a large external “star” button (i.e. “Go” stimuli), but to resist responding to a fourth shape (i.e. the “No-Go” stimulus). The shape designated as the “No-Go” stimulus was counterbalanced between participants. To generate a prepotent response, 75% of trials were “Go” trials requiring a button press, and 25% of trials were “No-Go” trials where the response should be withheld.
The maximum inter-stimulus interval (ISI) (i.e. from stimulus onset to stimulus onset) was 2500ms. At the start of each trial, a fixation cross would appear at the centre of the screen for 200ms. This was then replaced by the stimulus, which remained on-screen for 200ms. After the stimulus offset, participants had a further 1000ms to respond, at which point the trial automatically terminated. Stimulus presentation was followed by a 1100ms pause before the next trial commenced. An error tone (“bleep”) was played immediately if the child made an omission error (i.e. failed to respond on a “Go” trial), or a false alarm (i.e. pressed the star button on a “No-Go” trial). A positive feedback-noise (“ping”) was played if the participant made a correct response.
Procedure. Before starting the task, each child completed a warm up session to familiarize with the “Go” and “No-Go” stimuli. Training was terminated only when the child could correctly identify the required response for each shape. Children then completed a short practice block of eight trials containing all four stimuli presented in a fixed, but superficially random order. Then followed 144 experimental trials split into three 48-trial blocks, each separated by a short break. Stimulus presentation was randomised throughout each half block to avoid clustering of “No-Go” trials. The task (including training) lasted approximately ten minutes.
Four measures of task performance were obtained:
1. Number of false alarms (or commission errors): “No-Go” trials on which the button was pressed. This is the main measure of inhibitory control, with false alarms representing failure to inhibit the prepotent button-press response.
2. Number of hits: “Go” trials on which the child responded. This is not a main measure of inhibitory control performance, but indicates how reliably participants detect targets when present and suggest the strength of the prepotent response generated. Hit-rates are also used for calculations of task sensitivity (see below).
3. Task Sensitivity: Estimates of participants’ task sensitivity can be calculated using signal detection theory (A0) and probability estimates of False Alarms and Hits. This permits differentiation between participants who make fewer false alarms, but also fewer hits (poor task sensitivity), and those who make fewer false alarms despite a good hit rate (good task sensitivity). This is important because a low false alarm rate could be due to a generally low response rate (for both “Go” and “No Go” stimuli). Task sensitivity (A0) is a nonparametric measure which ranges from 0.5 (chance performance) to 1 (perfect sensitivity), and is calculated as follows (Grier, 1971):
A = 0.5 (H-FA) (1+H-FA) / [4H (1-FA)]
Where, H = probability (Hits), FA = probability (False Alarms).
4. Hit Trial Reaction Time (RT): Although not a measure of inhibitory control per se, this might indicate between-group differences in processing speed and/or task-strategy.
Dog-Pig Stroop Task.
Task Design. The stimuli used in this task were two simple line drawings of a dog and a pig (see Appendix 4). Stimuli were presented centrally on the computer screen, subtending approximately 6° vertically and 9° horizontally. Two experimental conditions, each containing 32 trials were administered. In the control (baseline) condition, children were simply instructed to say “dog” when they see the dog-image, and “pig” when they see a pig as quickly possible. In the Stroop (i.e. inhibition) condition, children were instructed to say "dog" to pig images, and "pig" to dog-images, as quickly and accurately as possible.
Children’s responses were recorded by an assistant during the task, and also audiotaped so that manuscripts could be subsequently checked by the experimenter. If a child made a mistake on a trial and then corrected themselves, their initial response was recorded. To estimate response latency on each trial, the experimenter would press a large external button as soon as the child made their initial response. Although this measure of reaction time is relatively crude, many of the children taking part would not have been testable with throat microphones which measure voice-onset. These technologies are highly sensitive to all sounds including subtle body movements, lip-smacks and vocalizations, reducing their reliability for use with participants who might have difficulty minimizing task-irrelevant movement or vocalizations. It is also notable that the additional error in reaction-time estimates induced by this method would be constant across groups.
On each trial, the stimulus remained centrally on-screen until a response had been registered (i.e. the response button had been pressed). If no response had been registered after 3000ms had elapsed, the trial automatically terminated, and the message “Too Slow” was presented for 500ms. Stimulus presentation was followed by a 2000ms pause (inter-trial interval) before the next trial commenced. The maximum ISI was thus 5500ms.
Procedure. All children completed the control condition first to provide a measure of baseline picture naming speed and accuracy[2]. After the control condition had been completed, children were presented with training slides to familiarise them with the Stroop naming procedure. After successfully completing the four practice trials, children would commence the 32-trial Stroop condition block. The task (including training) lasted approximately 7 minutes.
Flanker Task
Task Design. For this computer task, children were presented with two large arrow-shaped buttons – one pointing left and one pointing right. There were three experimental conditions: baseline, congruent, and incongruent. Children were asked to respond by pressing the arrow-button pointing the same way as the white target arrow, which was positioned centrally, subtending approximately 4° vertically and 6° horizontally. On baseline trials, the white target arrow was presented on its own. On congruent trials, the white target arrow was flanked by four red ‘distractor’ arrows pointing the same way as the target (e.g. ààààà). On incongruent trials, the white target arrow was flanked by four red ‘distractor’ arrows facing in the opposite direction to the target arrow (e.g. ßßàßß). It is thus only on incongruent trials that the distractors must be actively inhibited/suppressed for correct target identification.
The maximum ISI was 2900ms. A fixation cross would appear centrally on-screen for 200ms. This was then replaced by the stimulus (neutral, congruent or incongruent), which remained on-screen until a button-press had been registered. If no response had been registered after 1200ms had elapsed, the trial automatically terminated. An error-tone (“bleep”) was played if the participant pressed the wrong arrow-button . If the child failed to respond before the trial terminated, an error-tone was played and a “Too-Slow” message was briefly displayed. When the child responded correctly, a positive feedback-noise was given (a “ping”). There was a 1100ms pause (inter-trial interval) between trials.
Procedure. Each child first completed a series of familiarisation trials. This was followed by three blocks of 30 trials separated by a short break (90 trials in total). Each block contained ten baseline, ten congruent and ten distractor trials, which were distributed randomly. Error-rates and mean reaction times (RT) for neutral, congruent and incongruent trials were recorded.
[1] Although a cut-off of 30-points is typically used with younger children, a slightly lower cut-off score is thought to be more accurate for use with older children/adolescents (Mesibov et al., 1989). This is due to the inclusion of one or two items on which older children with autism tend not to score highly (e.g. imitation).
[2]Condition-order was fixed because a pilot study showed that if children completed the experimental (i.e. Stroop) condition first they had difficulty forgetting the ‘opposite’ rule in order to name the pictures normally for the control condition. This was shown by elevated error-rates and poorer naming speeds. Therefore, in order to obtain a realistic measure of ‘automatic’ (i.e. control-condition) picture naming speed and accuracy, and a stronger prepotent response, it was decided appropriate to fix the order of condition presentation (Control, then Stroop). Although this may lead to practice effects, this effect is constant across groups.
Publisher
Lancaster University
Identifier
sanderson2010
Contributor
John Towse
Rights
Open
Language
English
Type
project description
Coverage
LA1 4YF
LUSTRE
Supervisor
Melissa Allen
Project Level
MSc
Topic
Developmental Psychology
Sample Size
Autism group. Thirty-five individuals with autism, aged between 6 and 18 years
Control group. Thirty typically developing (TD) children, aged between 6 and 11 years, were recruited from three state primary schools
Control group. Thirty typically developing (TD) children, aged between 6 and 11 years, were recruited from three state primary schools
Statistical Analysis Type
ANOVA
MANOVA
Chi squared
correlation
MANOVA
Chi squared
correlation
Files
Collection
Citation
Charlotte Sanderson, “The Specificity of Inhibitory Impairments in Autism and Their Relation to ADHD-type Symptoms,” LUSTRE, accessed April 28, 2024, https://www.johnntowse.com/LUSTRE/items/show/21.