Preliminary evidence for an association between intake of high‐fat high‐sugar diet, variations in peripheral dopamine precursor availability and dopamine‐dependent cognition in humans

Obesity is associated with alterations in dopaminergic transmission and cognitive function. Rodent studies suggest that diets rich in saturated fat and refined sugars (HFS), as opposed to diets diets low in saturated fat and refined sugars (LFS), change the dopamine system independent of excessive body weight. However, the impact of HFS on the human brain has not been investigated. Here, we compared the effect of dietary dopamine depletion on dopamine‐dependent cognitive task performance between two groups differing in habitual intake of dietary fat and sugar. Specifically, we used a double‐blind within‐subject cross‐over design to compare the effect of acute phenylalanine/tyrosine depletion on a reinforcement learning and a working memory task, in two groups that are on opposite ends of the spectrum of self‐reported HFS intake (low vs high intake: LFS vs HFS group). We tested 31 healthy young women matched for body mass index (mostly normal weight to overweight) and IQ. Depletion of peripheral precursors of dopamine reduced the working memory specific performance on the operation span task in the LFS, but not in the HFS group (P = 0.016). Learning from positive‐ and negative‐reinforcement (probabilistic selection task) was increased in both diet groups after dopamine depletion (P = 0.049). As a secondary exploratory research question, we measured peripheral dopamine precursor availability (pDAP) at baseline as an estimate for central dopamine levels. The HFS group had a significantly higher pDAP at baseline compared to the LFS group (P = 0.025). Our data provide the first evidence indicating that the intake of HFS is associated with changes in dopamine precursor availability, which is suggestive of changes in central dopamine levels in humans. The observed associations are present in a sample of normal to overweight participants (ie, in the absence of obesity), suggesting that the consumption of a HFS might already be associated with altered behaviours. Alternatively, the effects of HFS diet and obesity might be independent.


| INTRODUC TI ON
Over recent decades, obesity has become a global health burden, making research on the development and maintenance of obesity more relevant than ever. One of the main drivers of the rapid rise in obesity rates is the increased intake of food products containing high amounts of saturated fat and refined sugars. 1 The question is, what makes people eat beyond their caloric needs, despite negative consequences, such as getting uncomfortably full or the health risks associated with obesity?
Throughout their daily life, people are constantly exposed to food advertisements and easily available food products. Such external food cues have the potential to enhance the motivation to obtain and consume food, even in a satiated state. 2 Recently, it has been shown that people with obesity outperform people with normal weight when learning and tracking the reward predicting value of cues associated with a food reward. 3 In addition, individuals with a higher body mass index (BMI) compared to a lower BMI (normal weight to obese) continue to respond to such food reward cues with the same intensity, despite their decreased motivation to consume the food rewards after devaluation 4,5 In a meta-analysis, García-García et al 6 showed that people with obesity exhibit hyperactivation in reward-related brain areas and proposed that this enhanced focus on rewards may lead to compulsive-like behaviours. In addition to motivational aspects and behavioural control, obesity is associated with altered decision making and executive functions. 7,8 Adverse decision making might be explained by the inability to integrate negative feedback as shown by impaired reinforcement learning associated with obesity. 9 In a probabilistic reinforcement learning paradigm with monetary rewards, people with obesity chose the correct option less frequently and gained lower overall payout compared to lean participants. 10 Coppin et al [10][11][12][13] report similar findings of not only impaired reinforcement learning in obesity, but also impairments of working memory, in line with previous findings. 11,12 The observed alterations of cognitive processes linked to motivation and behavioural control may contribute to the maintenance of obesity and are considered to be a result of alterations in central dopamine pathways. 7 Reinforcement learning and working memory both depend on action of dopamine in the striatum and prefrontal cortex (PFC) and optimal levels are crucial for proper functioning. [14][15][16] Although alterations in the dopaminergic system have mainly been associated with body weight in humans, [17][18][19][20][21][22] studies in rodents suggest that diets high in saturated fat and refined sugar (HFS), as opposed to diets low in saturated fat and refined sugars (LFS), lead to the observed changes, independent of excessive weight: exposure to high-fat diets reduced dopamine receptor D 2 protein expression levels, 23 affected dopamine synthesis 24,25 and uptake of striatal dopamine in rodents. 26,27 Furthermore, the overconsumption of specifically saturated dietary lipids, predominating in a typical western style diet, reduced dopamine receptor D 1 signalling in rats, as well as impaired dopamine clearance and phasic dopamine release in the nucleus accumbens in mice independent of weight gain. 28,29 Mimicking the effects of hidden sugars in commercial foods and beverages, low-concentration sucrose solutions changed dopamine receptor D 1 and D 2 mRNA and protein expression in the striatum. 30 Furthermore, a high-fat diet down-regulated the expression of striatal dopamine receptor D 1 and D 2 mRNA. 31 However, it is not clear whether the observed alterations of the dopaminergic system are directly caused by HFS or are compensatory adaptations in response to altered dopamine levels.
Taken together, HFS diets may thus be responsible for the observed differences in adaptive behaviour that crucially rely on the neurotransmitter dopamine and that promote the overconsumption of such food products and obesity. However, translating the findings obtained from animal studies to humans has to be carried out with great care because of the large knowledge gap between the fields. 32 To date, a possible relationship between HFS diets and the dopamine system has not been investigated in humans. Here, we aimed to obtain evidence indicating that a high (relative to low) dietary intake of saturated fat and free sugars is associated with alterations of central dopamine and dopamine-dependent cognition, particularly, reinforcement learning and working memory.
Because the synthesis of monoamine neurotransmitters in the brain depends on the availability of their amino acid precursors circulating in the blood [peripheral dopamine precursor availability (pDAP)], central dopamine levels can assumedly be decreased by depleting its precursors tyrosine and phenylalanine relative to the other large neutral amino acids, which competitively share a carrier at the blood-brain barrier. 33 To uncover potential diet-related differences in central dopamine levels and consequently dopamine-mediated cognition, we made use of an acute phenylalanine/ tyrosine depletion (APTD) method, which attenuates dopamine synthesis and release in the striatum [34][35][36] and impairs frontostriatal functional connectivity. 37 The effect of APTD on central dopamine synthesis and release has been shown in human positron emission tomography (PET) studies and is further substantiated by evidence from animal research. APTD increases baseline neuronal firing and amphetamine-induced [ 11 C]raclopride binding potential in the striatum, which has been interpreted as reduced dopamine release. 35,36,38 Applying APTD in rats revealed reduced tyrosine that the consumption of a HFS might already be associated with altered behaviours.
Alternatively, the effects of HFS diet and obesity might be independent.

K E Y W O R D S
acute phenylalanine/tyrosine depletion, dopamine, high fat and sugar diet, reinforcement learning, working memory levels in the striatum, frontal cortex and hippocampus, as well as reduced accumulation of the dopamine precursor dihydroxyphenylalanine (DOPA) (synthesised from tyrosine) in the same brain regions. 33 Taken together, the results from human and animal studies suggest that APTD decreases central dopamine synthesis and release. To what extent exactly central dopamine release is decreased by precursor depletion and whether this decrease is of similar magnitude between individuals is not known; higher concentrations of striatal dopamine or presynaptic dopamine synthesis capacity, as shown for women compared to men, 39,40 could serve as buffer for the effects of peripheral depletion. APTD has been shown to modulate reinforcement learning 41,42 and executive functions such as set-shifting and spatial working memory. 37,38,43 These cognitive processes require a certain level of dopamine for optimal performance. Either a decrease or increase in this level will lead to suboptimal performance (ie, dopamine levels relate to cognitive performance in an inverted-U-shaped manner). 16 As such, assessing reinforcement learning and working memory performance after APTD in two groups that differ markedly in their dietary intake of saturated fat and free sugars could reveal potential diet-related differences in the dopamine system.
Our main hypothesis for this study is that APTD differentially affects cognitive performance of two groups that differ in their self-reported intake of saturated fat and refined sugar, as a result of potential diet-associated alterations of the dopaminergic system. As secondary hypothesis, we expect markedly reduced pDAP levels after APTD in both diet groups, indicating effective intervention and causing the changes in cognitive performance analysed in our main hypothesis. Exploratory analyses aimed at characterising the two diet groups with respect to eating behaviour, personality traits, metabolic hormones, and parameters of fat and sugar metabolism.

| Participants
Ninety healthy female participants (mean ± SD age, 25.03 ± 3.61 years; BMI = 24.16 ± 5.72 kg m -2 ) were recruited from the internal participant database of the Max Planck Institute for Human Cognitive and Brain Sciences (Leipzig, Germany) and via advertisements placed at university facilities or public spaces. All participants were nonsmokers and reported no history of clinical drug or alcohol abuse or neurological disorder, and none had a first-degree relative history of psychiatric disorders. None showed moderate or severe depressive symptoms assessed with the Beck Depression Inventory (BDI), indicated by total scores < 19. 44 We decided to only include female participants because previous studies reported larger behavioural effects of APTD in women compared to men, 42,45 an effect potentially explained by higher striatal dopamine synthesis capacity in women. 40 The study was carried out in accordance with the Declaration of Helsinki and was approved by the Medical Faculty Ethics Committee of the University of Leipzig (439/16-ek). All participants provided their written informed consent before taking part in the study.

| Study design
Participants were first invited to the institute for a screening to check for inclusion and exclusion criteria. We used the Dietary Fat and free Sugar Questionnaire (DFS) 46,47 to group our participants into two groups of high and low consumers of saturated fat and refined sugar (HFS vs LFS group). The DFS consists of 24 questions asking how often participants consumed a certain food item on average over the last 12 months (five answer options; from "one per month or less" to "five times or more per week"). Two additional questions ask for the frequency of food consumed outside the home averaged over the last 12 months and the number of spoons of sugar added to food and beverages in the last week. The minimum score possible is 26 (low intake of HFS) and the maximum score is 130 (high intake of HFS). DFS scores were shown to correlate with the percentage energy from saturated fat and free sugar and high intra-class correlations indicate good test-retest reliability. 46,47 Based on the total DFS score, participants were assigned to the LFS group (total score < 54) or the HFS group (total score > 61); participants with DFS scores ranging from 54 to 61 were excluded from the study. Additionally, baseline fasted blood measurements were taken, and participants completed the Viennese Matrices Test 2 (WMT-2) to assess intelligence. 48 They further answered self-reported questionnaires on eating behaviour and personality.
If participants fulfilled all inclusion and none of the exclusion criteria, they underwent two test days with a minimum of 7 days between sessions (mean 11.97 days, maximum 36 days) (Figure 1).
A within-subject, double-blind, cross-over design was used to test participants under a dopamine depletion condition (DEP) and a balanced dopamine condition (BAL); the intervention was administered in balanced order. Test sessions were scheduled either at 08.00 am or 10.00 am, the two sessions always started at the same time for each participant. Before ingestion of the amino acid drink and at the end of the test session, participants rated their well-being with digital visual analogue scales (VAS) asking for sadness, anxiety, mood, nausea, appetite, hunger, satiety, fullness and urge to move.
To monitor success of the APTD intervention, blood samples were drawn before ingestion of the drink and approximately 4 hours post ingestion, prior to behavioural testing. To assess item span, which has been considered a proxy for dopamine, 16 the verbal forward and backward digit span task 49 was administered in a soundproof room immediately before behavioural testing. Behavioural testing was conducted 4.5-5 hours post ingestion (mean: 4 hours 49 minutes, maximum: 5 hours 38 minutes). During the period between ingestion and behavioural testing participants read, watched a movie or worked quietly. Two hours after ingestion participants were provided with a low protein snack, consisting of fruits (apple, banana and grapes) and vegetables (cucumber, carrots and red pepper).

| Behavioural testing
Participants performed the probabilistic selection task (PST) 15

| OSPAN
Working memory performance was examined with a modified version of the automated OSPAN task. 50,51 During the OSPAN task, participants had to mentally solve a presented mathematical problem (eg, [4 × 2] -7) and then indicate with a mouse click, whether the presented answer is the correct answer to that problem. The time limit for answering was the average time that participants needed to answer the given solutions to mathematical problems in the preceding training phase plus 2.5 SD. Subsequently, a target letter was presented on the screen, which participants were instructed to remember. After three to seven items (with the number of items per trial varying randomly to prevent participants from anticipating the number of items to be remembered), participants were asked to recall the items by choosing letters from a 3 × 4 matrix containing 12 letters and clicking them in the presented sequence with the mouse.
Each length of items was presented three times, adding up to a total of 75 math problems and letters presented. The complete task with training and test phase takes around 30 minutes to finish.
Working memory performance was calculated using the MIS scoring method, a measure that accounts for performance on the distractor task that we have developed previously. 52 In short, the MIS main score (referred to as "MIS score") consists of the working memory related components "number of remembered items" (I) (short-term memory) and "longest contiguous sequence remembered" (S) (relative object placement) and adjusts for performance on the mathematical distractor task (M) on each trial. The MIS score for each trial was calculated using: The left side of the multiplication accounts for performance on all mathematical problems (M) except the first one presented, by calculating the ratio of the number of correctly answered problems minus one to the total of mathematical problems minus 1 ( ). This part of the score is divided by two to weight the distractor and the working memory part of the score equally. The total MIS score for each participant is the sum of all scores per trial; the maximum MIS score possible is 15. The MIS scoring method allows to calculate a subscore only for the working memory components of the OSPAN without the distractor task by only calculating the right side of the multiplication shown above (referred to as the "IS subscore"). The total IS subscore for each participant is the sum of all scores per trial; the maximum IS subscore possible is 15.

| PST
The PST consists of a training and a test phase. 15

| APTD
To first deplete pDAP levels, participants followed a diet low in protein (< 20 g of protein) on the day prior to the test sessions (guidelines provided by a nutritionist) and fasted overnight from 10.00 pm.
Drinking water was encouraged and drinking black coffee and tea

| Self-reported questionnaires
All participants completed the BDI and DFS for inclusion, as well as personality and eating behaviour questionnaires to characterise the two diet groups on the screening day. Feeling of hunger, dietary restraint and disinhibition were assessed using the Three Factor Eating questionnaire (TFEQ). 55,56 Personality measures encompassed the personality traits openness, conscientiousness, extraversion, agreeableness and neuroticism (NEO-FFI), 57 behavioural inhibition and approach system (BIS/BAS) 58 and impulsivity (UPPS). 59

| Blood measures
Blood samples for the analyses of amino acids were drawn at the screening, as well as prior to ingestion of the drink and prior to be-

| Study samples
Sixty-five participants completed the screening day (ie, had not to be excluded based on health issues, smoking or drug abuse), including blood drawing and self-reported measures of eating behaviour and personality, and began the test days with dietary intervention.

F I G U R E 2
Flowchart of the study protocol. Enrolled participants were screened for eligibility based on health and diet. Included participants completed two test days varying in intervention drink. Participants with unsuccessful intervention (for details, see Materials and methods) or who vomited/felt nauseous during testing were excluded from the analyses. Task specific criteria were used to define samples for task analyses. Dashed blue frames indicate samples that were used for statistical analyses. BMI, body mass index; LFS, low fat sugar; HFS, high fat sugar During the course of the study, 16 participants dropped out voluntarily ( Figure 2). A further three participants with an estimated IQ lower than 85 and four participants who had to vomit after the ingestion of the amino acid drink on one of the test days were excluded from the analyses. Finally, 11 participants for whom the intervention was unsuccessful had to be excluded from the analyses. Statistical outliers for BMI, based on the 2.2 interquartile range, were included in the analyses to ensure proficient sample size. Thus, the study sample consisted of 31 subjects, 17 in the LFS group and 14 in the HFS group (Table 1) Because a secondary aim of the study was to characterise the two dietary groups with respect to the dopaminergic system, but also metabolic parameters, eating behaviour and personality, we

| Statistical analysis
Statistical analyses were performed in r, version 3.6.1 62 within rstudio, 63 using the packages car, stats, pastecs, lme4 and psych.
Group demographics (age, IQ and BMI) and questionnaire data were tested using Welch's t-test for unequal variance. Metabolic parameters were analysed with linear regression models using the function lm of the r package stats with diet group (LFS vs HFS) as predictor and BMI as covariate. All data that were measured multiple times (OSPAN and PST performance and reaction times, digit span [forward and backward], pDAP, VAS) were analysed with linear mixed-effect models with random intercepts and fixed slope for the random factor subject, using the function lmer of the r package lme4, unless stated differently. To test the significance of an effect in question, we compared the full model to a null model without the effect in question with likelihood ratio tests. 95% Confidence intervals (CI) are reported for significant effects to illustrate certainty of this effect (confidence levels displayed as 0.000 are numbers smaller/ larger than zero, indicating the CIs do not cross zero). The assumption of normally distributed residuals of the models was checked by visually inspecting the qq-plots and no obvious violations were found. Primary performance measure for the cognitive tasks (OSPAN and PST) was accuracy, secondary performance measure were reaction times (RTs). OSPAN performance was analysed in a model with fixed effects diet group (LFS vs HFS) and intervention (BAL vs DEP) and the diet group × intervention two-way interaction, controlled for BMI and test day to account for training effects.
Performance in the PST training phase was analysed in an ordinal regression model using the function clmm of the R package ordinal, with fixed effects diet group (LFS vs HFS) and intervention (BAL vs DEP), and the diet group × intervention two-way interaction, controlled for BMI and test day; number of learn blocks were ranked from lowest to highest (1-10) and not reaching the test phase was assigned the highest rank 11. Performance in the PST test phase was analysed in a model with fixed effects diet group (LFS vs HFS), intervention (BAL vs DEP) and task condition (approach vs avoid), the diet group × intervention × task condition three-way interac-

| RE SULTS
The present study was designed to investigate the differential ef-

| Working memory
On each test day, we measured simple item span with the forward and backward digit span task to account for possible effects of the intervention on diet groups and group differences in short-term memory that might explain different performance on the OSPAN

| Reinforcement learning
Reinforcement learning was tested with the PST, which consists of a training and a test phase. 15 Learning of reward associations during the training phase did not differ between diet groups (χ 2 = 1.26, df =1, P = 0.262) or interventions (χ 2 < 0.01, df =1, P = 0.952). The diet group × intervention interaction was also nonsignificant (χ 2 < 0.01, df =1, P = 0.982). Reaction times in the training phase did not differ between diet groups, interventions and there was no significant diet group × intervention interaction (all P > 0.488). The test phase of the PST tests how well participants learned to approach rewarded stimuli and avoid punished stimuli (referred to as task condition). Analysis of the PST test phase revealed no significant diet group × intervention × task The main effect of intervention reached significance (χ 2 = 3.88, df =1, P = 0.049, 95% CI = 0.000-0.132), indicating that APTD increased accuracy of approach as well as avoid choices in both diet groups ( Figure 5A). Analyses of reaction times in the test phase revealed no significant diet group × intervention × task condition interaction (χ2 = 0.08, df =1, P = 0.771). The diet group × intervention interaction was significant (χ2 = 15.11, df =1, P < 0.001, 95% CI = 188.64-541. 86), indicating that the HFS group responded faster on the BAL than the DEP day in both task conditions, whereas reaction times were similar for the LFS group on both test days (post-hoc Tukey's test, BAL HFS vs DEP HFS , t 73.2 = −4.65, P < 0.001) ( Figure 5C).

| The effects of dopamine-depletion on mood and well-being
To check whether dopamine depletion affects potential confounders such as mood and well-being differentially in the two diet groups,

| pDAP
We assessed group differences in pDAP at baseline (screening, prior ingestion of amino acid drinks on test days) as a proxy for the sta-

| Self-reported eating behaviour and personality traits
Because the preference for HFS might be influenced by general differences in eating behaviour, we investigated potential group differences on the three factor eating questionnaire. The HFS group The two diet groups did not show significant differences on any of the subscales of the NEO-FFI, BIS/BAS or UPPS questionnaire (all P > 0.08).

| Metabolic blood parameters
We analysed parameters of the fat and sugar metabolism and eating related hormones to check if the dietary preference of the groups is reflected in physiological measurements. The two dietary groups did not differ in any parameter of fat (cholesterol and triglycerides) or sugar metabolism (glucose and HbA1c), as well as leptin, insulin and insulin resistance (all P > 0.503) ( Table 1).

| Further characterisation of diet groups (remaining sample)
Because an aim of the present study was to characterise the two dietary groups with respect to the dopaminergic system, but also metabolic parameters, eating behaviour and personality, we repeated the analyses of measurements obtained at screening day with the remaining participants that completed the screening day and were not excluded based on health issues, but were also not eligible for the main sample (see Supporting information, Table S1).
The two groups in the remaining sample did not differ in age

| The effect of pDAP on dopaminedependent cognition
The two diet groups differed significantly in pDAP at screening ( Figure 6), when measurements should reflect pDAP levels associated with participants' regular diet. Because pDAP can be considered a proxy for central dopamine release 35,36,38 and chronically higher release of dopamine might induce adaptive changes in the dopaminergic system, such as higher sensitivity of receptors, 64 we included pDAP at screening as predictor in our models for cognitive performance (OSPAN and PST) and simple span (digit span) instead of diet group. ter under the DEP than the BAL condition ( Figure 5B). In the avoid condition, participants with lower pDAP performed worse under the BAL than the DEP condition and performance was unchanged for participants with higher pDAP. Analyses of reaction times in the test phase revealed no significant pDAP × intervention × task condition interaction or any of the lower two-way interactions and no significant main effect of pDAP (all P > 0.118) ( Figure 5D). The DEP intervention increased reaction times irrespective of pDAP and task condition (main effect of intervention, χ 2 = 5.29, df =1, P = 0.021, 95% CI = 17.66-214.33) and all participants responded slower in the avoid than the approach condition (main effect of task condition, χ 2 = 5.05, df =1, P = 0.028, 95% CI = 14.78-209.77).

| D ISCUSS I ON
We aimed to provide first evidence indicating that habitual dietary intake of saturated fat and added sugar is associated with altera-   before: in line with our finding, higher neuroticism is associated with higher preference for and consumption of sweet foods. 73,74 However, it is still debatable whether personality traits influence food consumption 75 or whether more basal factors such as genetic predisposition are stronger contributors. 76 Note that the dopamine depletion effects have to be interpreted with care and await future replication. It also should be considered that some of the findings in the main sample could not be replicated in the second sample of screened participants.
Specifically, the higher pDAP observed in the in HFS group of the main sample has to be interpreted with care because this finding was not significant in the remaining screening sample. We also could not replicate the differences in higher disinhibition in eating behaviour in this second screening sample. This calls for replication in different study populations. The small sample size is a major limitation of our study and the results have to be interpreted with caution because of the concomitant low statistical power. In consequence, our findings have to be considered as preliminary, requiring replication with a higher sample size to provide sufficient statistical power for detecting smaller effects, which have to be assumed when studying diet. The size of the main sample was low due to an unusually high dropout rate (Figure 1) compared to other studies that administered APTD. 36,41 Nausea or vomiting is a common side-effect of ATPD, likely because of its unpalatable taste 42,45,77 ; however, our administration of the APTD intervention differed in the sense that we mixed the amino acid drink with lemonade instead of syrup and also that syrup might have a stronger flavor to disguise the bitter taste of amino acids. Furthermore, other studies administered the unpleasant amino acids such as methionine separately from the dissolved mixture 78,79 to reduce risk of nausea. We recommend that future APTD studies follow these precautions. Additionally, we recognised that female participants on average report more nausea or regurgitating, in contrast to male participants. 36,54,78,80 Furthermore, the generalisability of our findings is limited because we only included young healthy women in this study. Dopamine availability in the striatum appears to depend on gender 81 and cortical plasticity is influenced by levels of sex hormones, 82 factors that might determine the strength of putative diet-induced changes of the dopaminergic system.
Additionally, it should be noted that, unfortunately, we were not able to control analyses of the cognitive tasks for menstrual cycle as a result of missing cycle data for some of the participants because the levels of the sex hormone oestradiol have been shown to affect dopamine-dependent cognition such as working memory and reinforcement learning. [83][84][85] We are also not able to make any statement about possible interaction effects of HFS and obesity because our sample mainly included participants from the normal to overweight range. We are also aware of the fact that the genetic background influences baseline dopamine transmission parameters and cognitive function, 86 which we cannot account for in our study. Future studies, including men and women, focusing on a more narrow range of BMI and with a sample size sufficiently large to consider genotypic variation affecting dopaminergic transmission, are needed to shed further light on the association of HFS and the dopaminergic system in humans.

| CON CLUS IONS
The present study provides the first evidence indicating that the amount of saturated fat and refined sugars habitually consumed is associated with the different availability of dopamine precursors in humans that could potentially explain differential effects of a dietary dopamine manipulation. We provide first evidence indicating that (i) the effect of a dietary dopamine depletion on working memory (but not reinforcement learning) performance and (ii) peripheral availability of dopamine precursors, a proxy for central dopamine release, 35,36 differed between two groups reporting high relative to low intake of high fat and sugar food products. It has to be stated explicitly, however, that any conclusions drawn from the present study are limited by the low sample size and statistical power and thus await future replication.

ACK N OWLED G EM ENTS
We thank Arno Villringer and his coworkers of the Department We thank Frauke Beyer for support with the statistical linear and mixed models. We thank Lydia Hellrung for helpful input regarding the APTD procedure. This work was funded by the Deutsche

CO N FLI C T O F I NTE R E S T S
The authors declare that they have no conflicts of interest. Writing -review & editing.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/jne.12917.

DATA AVA I L A B I L I T Y
The data that support the findings of this study are available from the corresponding author upon reasonable request.