Construct	Instrument	Psychometrics	Description
Intelligence	Wechsler Abbreviated Scale of Intelligence, second version, Dutch edition (WASI-II NL)	The Dutch version of the WASI-II NL has not been validated yet, but the Dutch version of the WASI has been previously used in autism research.[@305637]	The WASI-II NL consists of two verbal (Vocabulary and Similarities) and two non-verbal tasks (Block design and Matrix Reasoning). For a description of the subtests of the WASI-II we refer to the WASI-II manual.[@305638] Items are scores on a two (0/1) or three point scale (0/1/2). The raw scores of every subtest are converted to standard scores and thereafter converted to scale and total scores.
Attention bias	Dot probe task (DPT)	The DPT has shown to be sensitive to detect attentional biases.[@305639] The psychometric properties of the paradigm for the AN group with the use of food stimuli have not been validated yet.[@305640]	Both our DPTs consist of two blocks with 60 trials. The paradigm starts with a fixation cross after which a pair with two stimuli (one of which is a food image and the other is a neutral stimulus, similar in shape and color) follows. Then, an asterix (i.e. probe) appears at the location of one of these stimuli. The participant has to press the left or right button on a two-button response box to indicate the location of the asterix. Reaction time and accuracy are measured for every trial. Differences between incongruent and congruent trial are studied.
Visual spatial abilities	Rey Complex Figure Test (RCFT)	The RCFT has good psychometric properties for the standard scoring systems[@305641] as proposed by Osterrieth[@305642]	The participant has to draw a complex figure, first by copying the figure from an example and then after 3 and 30 minutes from memory. Lastly, the participant performs a recognition task in which she has to indicate which elements among alternatives were shown in the original figure. Every element is scored on a 4-point (2/1/0.5/0). For the recognition trial scoring takes place on a 2-point scale (0/1). Raw scores are converted to standard scores. Additionally, we apply a slightly modified version of the Booth scoring method[@305643] as formerly used by Lang and colleagues[@305644] to obtain measures of local versus global processing.
	Motor-free Visual Perception Task- fourth edition (MVPT-4)	The MVPT-4 is a valid and reliable instrument to assess visuo-spatial abilities without using a motor response[@305645]	The test consists of different tasks in which the respondent as to point to the correct answer among alternatives. All of the 45 items are scored as true[@305566] or false (0). We gain measures of visual discrimination abilities, the ability to distinguish an object from its background (visual figure ground), visual memory skills, visual closure abilities and visual spatial relationships. Also a total raw score is yielded. Ray scores are converted to standard scores and an age equivalent.
	Navon task	The Navon task is a widely used task to assess global versus local processing, although the construct validity is questioned since local-global visual processing is not supposed to be a unitary construct.[@305646] Therefore the interpretation of the results should be done with caution	The participant looks at figures in the shape of a letter, which is build up from small letters. The participant has to indicate by pressing a key on a laptop whether an H or an O is presented in the stimulus, which demands the participant to examine the figure globally and locally. The stimuli is shown until response. Measures of accuracy and reaction times are obtained per trial. Differences between consistent and inconsistent stimuli are studied.
Set shifting skills	Set shifting neutral	The psychometric properties of the used paradigms have not been assessed yet, but the set up of our paradigms is similar to the category switch paradigm of Wolf and colleagues.[@305647]	Each paradigm has two blocks that consists of 32 trials. In the first version we use neutral stimuli and in the second version stimuli of sports/leisure activities are shown The neutral and sports/leisure activity stimuli are collected via Google search. In the third paradigm high caloric and low caloric food stimuli are shown. The stimuli are obtained from Blechtert’s food image database.[@305648] The neutral and activity paradigms are executed on a laptop and the food paradigm is performed during fMRI scanning. All paradigms start with a cue that indicates what task has to be executed. Thereafter a stimulus is presented and the task, indicated by the cue, has to be executed as quickly as possible. When the cue changes, it demands from the participant to switch strategies. The outcome measure for each task is the switch cost, calculated as the inverse efficiency scores (IESs[@305649; @305650]) by dividing the mean response time (RT) of correct responses by the proportion of correct responses (RT/[1 – ER]). Thus, participants with lower shifting competencies obtained higher scores in these tasks.
	Set shifting food
	Set shifting active
Inhibition skills	Go no go-neutral	The go no go paradigm is a widely used paradigm to assess inhibition skills. The psychometric properties of these particular paradigms have not been assessed yet, but the set up is similar to the go no go task as used by Wolf and colleagues.[@305647]	Each paradigm contains a go-block, consisting of 120 trials and a no-go block that also consists of 120 trials. In the first paradigm (neutral version, executed on a laptop) the participant has to press the button when dots are presented horizontally; when the dots are placed vertically they have to withhold their response. In the second and third paradigm, we used images of bodies, developed by Mousally and colleagues,[@305651] with a normal weight (second version, executed on a laptop) and bodies with a very low and very high weight (third version, executed during fMRI scanning). The instruction for both paradigms is to press the button on the response box when the body is in standing position. The participants have to withhold their response when the body is in a sitting position. The outcome measure for these tasks are the IES, calculated as the mean RT for correct go trials divided by the proportion of correct responses on no-go trials. Thus, participants with lower inhibiting capacities obtain higher scores.
	Go no go-normal
	Go no go-highlow
Decision making	Balloon Analogue Risk Task (BART)	The BART evidenced sound experimental properties and is supposed to be a useful tool in the assessment of risk taking .[@305652]	In this task the participant has to pump a balloon by pressing a button on the keyboard. With each pump the participant earns points. However, when the balloon collapses, the participant loses all points. This set up and analysis of the paradigm is described extensively elsewhere.[@305653] Reaction time and number of pumps per trial are obtained per trial. For a detailed description of the analysis of the BART we refer to Pleskac and Wershbale’s paper[@305653]
	Iowa Gambling Task (IGT)	Evidence provides support for the use of the IGT to detect decision making deficits in clinical populations, although data regarding reliability of the IGT are lacking[@305654]	The participant needs to choose one out of four card decks. The participant wins or loses money with each card they pick. Two out of four desks yield little money, but the participant doesn’t lose a lot of money either. The other two desks yield lots of money, but the participant is at risk to lose a lot of money as well. The number of choices from each deck during each trial are added up. We also calculate the number of “good” and “bad” desks .
	Probabilistic reversal learning paradigm (PRLT)	Probabilistic reversal learning tasks have been applied in AN studies previously e.g.[@305655; @305656] The paradigm as used in BRAVE has previously been used by Hooper[@305657]	The participant has to choose between a yellow or a blue square. Every time the participant receives feedback about whether this choice was right or wrong. The participant is instructed to gain as many right responses as possible. This PRLT has three conditions: in the first condition the participant receives positive feedback in 90% of the cases, in the second condition this was 80% of the cases and in the third condition 100% of the cases. The average number of trials to criterion (defined as the total trials for a discrimination or reversal minus the randomly determined criterion) is calculated separately for the different conditions. Errors (choosing the current “bad” color) are coded as: failures to maintain set, perseverative, or other.
Passive information processing	Passive task	As these paradigms are newly developed for this study, the psychometric properties are not evaluated yet.	The fMRI data analyses procedures are described elsewhere.[@305658] The eyetracking procedure has been described in Table 1.
	Free viewing
Executive functioning	The Behavior Rating Inventory of Executive Function (BRIEF)	The internal consistency of the Dutch BRIEF has appeared to be very high and has a high test-retest stability. The Dutch BRIEF is supposed to be a reliable measure of executive functioning.[@305659]	The 86 items of the BRIEF are answered on a 5 point Likert scale. The questionnaire provides a total score and two index scores (Behavioral regulation and Metacognition), which are derived from the following eight subscales: Inhibition, Shifting, Emotional control, Initiation, Working memory, Planning and organization, Organization of materials and Monitoring. Raw scores are converted to standardized subscale scores, index scores and a total score.