DJ-CP-M4

Vol.:(0123456789)1 3

European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071 https://doi.org/10.1007/s00406-019-01071-4

O R I G I N A L PA P E R

Impact of past experiences on decision‑making in autism spectrum disorder

Junya Fujino1,2 · Shisei Tei1,2,3,4 · Takashi Itahashi1 · Yuta Y. Aoki1 · Haruhisa Ohta1,5 · Manabu Kubota1,2,6 · Ryu‑ichiro Hashimoto1,7 · Motoaki Nakamura1,8 · Nobumasa Kato1 · Hidehiko Takahashi1,2,9

Received: 19 June 2019 / Accepted: 18 September 2019 / Published online: 26 September 2019 © Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract People are often influenced by past costs in their current decision-making, thus succumbing to a well-known bias recognized as the sunk cost effect. A recent study showed that the sunk cost effect is attenuated in individuals with autism spectrum disorder (ASD). However, the study only addressed one situation of utilization decision by focusing on the choice between similar attractive alternatives with different levels of sunk costs. Thus, it remains unclear how individuals with ASD behave under sunk costs in different types of decision situations, particularly progress decisions, in which the decision-maker allo- cates additional resources to an initially chosen alternative. The sunk cost effect in progress decisions was estimated using an economic task designed to assess the effect of the past investments on current decision-making. Twenty-four individuals with ASD and 21 age-, sex-, smoking status-, education-, and intelligence quotient-level-matched typical development (TD) subjects were evaluated. The TD participants were more willing to make the second incremental investment if a previous investment was made, indicating that their decisions were influenced by sunk costs. However, unlike the TD group, the rates of investments were not significantly increased after prior investments in the ASD group. The results agree with the previ- ous evidence of a reduced sensitivity to context stimuli in individuals with ASD and help us obtain a broader picture of the impact of sunk costs on their decision-making. Our findings will contribute to a better understanding of ASD and may be useful in addressing practical implications of their socioeconomic behavior.

Keywords Autism spectrum disorder · Decision-making · Sunk cost effect · Behavioral economics

Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0040 6-019-01071 -4) contains supplementary material, which is available to authorized users.

* Junya Fujino [email protected]

1 Medical Institute of Developmental Disabilities Research, Showa University, 6-11-11 Kita-karasuyama, Setagaya-ku, Tokyo 157-8577, Japan

2 Department of Psychiatry, Graduate School of Medicine, Kyoto University, 54 Shogoin-Kawaracho, Sakyo-ku, Kyoto, Japan

3 Institute of Applied Brain Sciences, Waseda University, 2-579-15 Mikajima, Tokorozawa, Saitama, Japan

4 School of Human and Social Sciences, Tokyo International University, 2509 Matoba, Kawagoe, Saitama, Japan

5 Department of Psychiatry, School of Medicine, Showa University, 6-11-11 Kita-karasuyama, Setagaya-ku, Tokyo, Japan

6 Department of Functional Brain Imaging, National Institute of Radiological Sciences, National Institutes for Quantum and Radiological Science and Technology, 4-9-1 Anagawa, Inage-ku, Chiba, Japan

7 Department of Language Sciences, Graduate School of Humanities, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo, Japan

8 Kanagawa Psychiatric Center, 2-5-1 Serigaya, Yokohama, Kanagawa, Japan

9 Department of Psychiatry and Behavioral Sciences, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo, Japan

1064 European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

Introduction

Decision-making is central to daily economic and social functioning [1–6]. Individuals with autism spectrum dis- order (ASD) frequently show altered decision-making in many situations [7–12], which can influence their socioec- onomic functioning [13, 14]. Improving our understanding of this issue could offer useful information for developing practical interventions to improve quality of life in ASD individuals.

Recently, studies on decision making and behavioral economics have rapidly expanded and efforts have been made to assess behavioral problems observed in patients with psychiatric disorders. These findings have begun to lay groundwork for improving the diagnosis and treat- ments for various psychiatric disorders [15–17]. Therefore, applying behavioral economics tools can help us to better understand altered decision-making abilities in ASD.

According to an assumption of traditional economics theory, decisions should be made based on the costs and benefits expected to arise in the future from the available options [18–22]. However, if costs of time, money, or effort were previously incurred, people frequently continue an investment or take an action, even though it has higher future costs than benefits [23–27]. This decision bias of considering past costs that cannot be recovered in current decision-making is referred to as the “sunk cost effect.” For example, imagine you bought a ticket for a basketball game. A few days later, your friend invites you to a spe- cial Italian dinner the evening of the game. Although you would prefer to attend the dinner, your thoughts revolve around the sunk cost of the already-paid-for ticket, and you decide to attend the game [18, 26]. Examples of the sunk cost effect also include decision-making related to invest- ments, such as the development of the supersonic plane Concorde. In early development stages, the plane was already significantly more expensive than expected, and the financial success of the project was unclear. However, the project was continued, and new funds were allocated to finish the plane to avoid wasting the significant amount of money already invested [18, 26].

The sunk cost effect has long been studied in various disciplines, including economics, psychology, politics, organizational behavior, and biology. Although the situa- tions in which the sunk cost effect is observed are diverse, they are generally divided into two types of decisions: utilization and progress decisions [26, 28]. As our first example illustrates, a utilization decision focuses on a decision maker confronted with the choice between two similar attractive alternatives, such that preferences shift to the sunk cost alternative. By contrast, our second exam- ple highlights a series of progress decisions in which the

decision maker allocates additional resources to an ini- tially chosen alternative; sunk costs increase the likelihood of further fund allocation. Because the influence of sunk costs can vary between utilization and progress decisions, previous studies call for a distinction between these two types of decisions when elaborating on the sunk cost effect [26, 28]. Please see Supplementary Introduction for details regarding prior research concerning the sunk cost effect.

The sunk cost effect is pervasive in real life and influ- ences many types of socioeconomic behaviors, such as investments [19], entertainment [26], management [29], and interpersonal relations [18]. Therefore, understanding how individuals with ASD behave under sunk costs may add crucial insights into the practical implications of their socioeconomic behaviors. However, to our best knowledge, only one recent study examined the sunk cost effect in indi- viduals with ASD, whose results suggested that the sunk cost effect is attenuated in ASD [30]. However, that study only addressed one situation of utilization decision focusing on the choice between similar attractive alternatives with dif- ferent levels of sunk costs. Thus, how sunk costs influence current decision-making in individuals with ASD remains largely unknown, particularly in progress decisions.

We, therefore, extended previous research by examining how individuals with ASD behave under sunk costs in a progress decision situation. Several studies have investigated various altered decision-making abilities in individuals with ASD. For example, the previous studies have examined the framing effect, a well-known decision bias, in which people differently react to a particular choice on monetary deci- sions depending on whether it is presented in a positive or a negative context [31, 32]. Compared with neurotypical participants, adults with ASD demonstrate less susceptibility to the framing effect and make more consistent choices [31, 32]. A recent study compared a series of choices regarding consumer products between individuals with ASD and con- trols and showed that the participants’ preferences between a given pair of options frequently switched when the third item in the set was changed [13]. This tendency decreased among individuals with ASD, thereby indicating more con- sistent and conventionally rational choices [13]. These previ- ous findings agree that individuals with ASD show reduced sensitivity to context stimuli and make more rational and consistent choices in experimental situations. Accordingly, we hypothesized that individuals with ASD would be less influenced by sunk costs under progress decision situations and would make more consistent and conventionally rational choices compared with individuals with typical development (TD).

1065European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

Methods

Participants

Twenty-five adults with ASD and 24 with TD were enrolled in this study. The sample size was determined on the basis of the previous studies assessing the deci- sion-making abilities of individuals with ASD [32–34]. Participants with ASD were recruited from a database of volunteers with an established clinical diagnosis of ASD in the outpatient units of the Showa University Karasuy- ama Hospital. The diagnostic procedure to identify indi- viduals with ASD was the same as in our previous studies [35–37]. Briefly, at least three experienced psychiatrists and a clinical psychologist assessed all the patients using the criteria of the Diagnostic and Statistical Manual of Mental Disorders, fourth edition text revision (DSM-IV- TR). The assessment comprised patient interviews about developmental history, present illness, life history, and family history. Patients were also asked to come accom- panied by suitable informants who had known them in early childhood. This process required approximately 3 h. A diagnosis of ASD was established only when there was a consensus between the psychiatrists and clinical psycholo- gist. At the time of testing, an experienced psychiatrist evaluated psychiatric comorbidities using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). No participants with ASD fulfilled the diagnostic criteria for substance use disorder, bipolar disorder, or schizophre- nia. Individuals with TD were recruited through advertise- ments and acquaintances. They did not meet the criteria for any psychiatric disorders as evaluated by an experi- enced psychiatrist using SCID. The ASD and TD groups were matched for age, sex, current smoking status, educa- tion, and estimated full-scale intelligence quotient (IQ) level. Smoking status is reportedly associated with various types of decision making [38]. No participants (ASD or TD) had any history of head trauma, serious medical or surgical illness, or substance abuse.

The ASD symptom severity and the IQ levels of the par- ticipants with ASD had been evaluated before the study. The Autism Diagnostic Observation Schedule (ADOS) [39] was used to assess ASD symptom severity. The IQ levels of the participants with ASD were evaluated using either the Wechsler Adult Intelligence Scale-Third Edi- tion (WAIS-III) or the Wechsler Adult Intelligence Scale- Revised (WAIS-R). Although the WAIS-III and WAIS- R have minor differences (e.g., more items), the number of core items largely remains unchanged. Therefore, we considered them essentially identical for the full-scale IQ measurement of individuals with ASD. Each ASD partici- pant was considered high functioning, because his or her

full-scale IQ score was above 80. The IQ scores of the TD participants were estimated using a Japanese version of the National Adult Reading Test (JART) based on the previous findings that the JART successfully predicted full-scale IQ scores in healthy populations [40, 41].

This study was approved by the institutional review board of Showa University Karasuyama Hospital and was con- ducted in accordance with the Code of Ethics of the World Medical Association. After providing a complete study description to all participants, written informed consent was obtained from all participants.

Economic task

We modified a clear example of the sunk cost effect in pro- gress decisions [19, 21, 26]. The current task consisted of 216 trials. During each trial, the participants were presented a project characterized by its costs and probability of success (Fig. 1). The project costs were ¥100 (approximately $1) or ¥275, and the probability of success was 40%, 50%, or 60%. These probabilities of success were actually implemented in the program. The participants had 5 s to decide whether to invest by pressing the corresponding key. The location of the “Invest” and “No invest” responses on the screen was counterbalanced during the task. If the participants did not respond within 5 s or decided not to invest, the trial was aborted. However, if they decided to invest, they received either the immediate feedback of whether the project was successful (control condition), or they were informed that further investments were required (sunk cost condition). In the latter case, the participants were next shown the addi- tional costs required and the current probability of success. The additional costs were again ¥100 or ¥275, and the prob- ability of success was again 40%, 50%, or 60%. The only difference between the decision scenarios for the initial investment and the follow-up investment was whether the participants had already invested in the project. Again, the participants had 5 s to decide whether to invest the addi- tional costs or stop the project. If the participants invested the additional costs, they received immediate feedback on the project’s success (i.e., there was at most one follow-up investment). If the participants decided not to invest the additional costs, the trial was aborted. The probabilities in the initial and follow-up decision scenarios were independ- ent, and any initial investment was lost, irrespective of the follow-up decision.

All six trial types that resulted from the different com- binations of project costs (¥100 or ¥275) and probabilities of success (40%, 50%, or 60%) were equally presented (36 times each). To ensure that there was a sufficient number of trials in which the influence of prior investments on cur- rent investment decisions could be tested (i.e., in which par- ticipants had decided to invest), two-thirds of all trials (24

1066 European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

trials) were “sunk cost conditions.” The trial order was rand- omized across participants. The participants could rest three times during the task. The time course is shown in Fig. 1.

The participants gained ¥1000 for each project that was completed successfully. However, they also had to pay for the investments they made in a trial, irrespective of a pro- ject’s success. We explained that their winnings were defined according to the outcomes of three trials after they had fin- ished (after the task, we debriefed them on the purpose of the experiment and paid a maximum predefined participa- tion fee), based on the previous studies [14, 42].

Before beginning, we performed a numeracy test to assess the participants’ numeracy skills and understanding of numbers (Supplementary Methods). All participants were quizzed regarding how well they understood the task (Sup- plementary Methods) and practiced a short version of the task at least once. Only after they successfully completing the quiz were they allowed to proceed to the experiment. The experiment was presented using E-Prime software (Psychol- ogy Software Tools, Inc., Pittsburgh, PA, USA).

Statistical analysis

First, the rates of investments (dependent variables) were submitted to a 2 [group (TD vs. ASD)] × 2 [condition (con- trol vs. sunk cost)] mixed analysis of variance (ANOVA).

Next, based on the amount of prior investments, the tri- als under the sunk cost condition were further subdivided into those in which the participants had already invested ¥100 (low sunk cost trials) or ¥275 (high sunk cost trials). Then, we estimated a sunk cost effect score for each partici- pant based on the previous studies [19, 21]. Because there was a significant intergroup difference in overall rates of investments in the control condition (TD 0.56 ± 0.17, ASD 0.79 ± 0.18, p < 0.01), we adjusted the rates of investments across groups by dividing rates of investments for each of six combinations of project costs and probability of success in three trial types (control, low sunk cost, and high sunk cost) by overall rates of investments in the control condition. Subsequently, we calculated individual differences in the rates of investments between “control trials” and “low sunk

Fig. 1 Experimental design. In each trial, the participants were pre- sented a project that was characterized by its costs and probability of success. The participants were instructed to decide whether they wanted to invest the requested amount of money in the project. If they decided to invest in the project, they received either the immediate feedback that the project was successful or not (control condition), or

they were informed that further investments would be required (sunk cost condition). In the latter case, the participants were next shown the additional costs that would be required and the current probability of success. Subsequently, the participants were instructed to decide whether to invest the additional costs or stop the project

1067European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

cost trials” and the difference between “low sunk cost tri- als” and “high sunk cost trials” for each of six combinations of project costs and probabilities of success. The average difference was used as an indicator of the strength of each participant’s sunk cost effect. A high sunk cost effect score indicates a stronger sunk cost effect.

We performed correlation analyses between the sunk cost effect score and the severity of clinical symptoms evaluated using the ADOS (communication subscale score, social interaction subscale score, and communication sub- scale + social interaction subscale score) across the partici- pants with ASD. Statistical analyses were performed using SPSS 24. Results were considered statistically significant at p < 0.05 (two-tailed).

Results

Three TD participants and one participant with ASD were excluded from the analyses (see Supplementary Results for details). Thus, data from 21 TD participants and 24 with ASD were analyzed. Demographic and clinical data are shown in Table  1. There were no significant differences between the groups in age, sex, current smoking status, edu- cation, and estimated full-scale IQ levels.

Overall, the participants performed the task well, miss- ing an average of only 0.58 ± 1.23 (mean ± SD) trials (total, 216 trials). In a missed trial, the participant could not make a decision within the time allowed (5 s, Fig. 1). There was no significant difference between the groups in the number of missed trials (p = 0.37).

Figure 2 shows the rates of investments in the sunk cost condition and the control condition in both groups. A 2 × 2 mixed ANOVA revealed the main effects of both group

(F = 17.55, p < 0.01) and condition (F = 11.90, p < 0.01). We also observed a significant group × condition inter- action (F = 4.49, p = 0.040), indicating that the TD and the ASD groups displayed differing patterns of decision- making. The TD group showed increased rates of invest- ments in the sunk cost condition compared with the con- trol condition (p < 0.01); no such difference was observed

Table 1 Demographic and clinical characteristics of participants

ADOS autism diagnostic observation schedule, ASD autism spectrum disorder, IQ intelligence quotient, TD typical development a Two-sampled t test b Two-tailed Chi-square test c Data not available for two participants

TD group ASD group Statistics (n = 21) (n = 24) p

Age (years) 32.7 ± 8.1 29.1 ± 3.7 0.07a

Male/female 18/3 22/2 0.53b

Current smoker/nonsmoker 5/16 3/21 0.32b

Education (years) 15.0 ± 1.7 15.4 ± 2.0 0.46a

Estimated full-scale IQ 106.4 ± 7.7 106.6 ± 13.1 0.96a

ADOSc

 Communication 3.9 ± 1.3  Social interaction 7.2 ± 2.1  Communication + social interaction 11.0 ± 3.1

Fig. 2 Rates of investments in the control and sunk cost condi- tions. The typical development (TD) group showed increased rates of investments in the sunk cost condition compared with the control condition (p < 0.01), whereas no such difference was observed in the autism spectrum disorder (ASD) group (p = 0.34). The error bars indicate ± standard errors. **p < 0.01

1068 European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

in the ASD group (p = 0.34). Reaction time had the main effects of condition (F = 13.22, p < 0.01), indicating that participants responded significantly slower in the sunk cost condition than in the control condition (sunk cost 1.20 ± 0.30 s, control 1.11 ± 0.26 s, p < 0.01). We found no significant main effects of group (F = 0.36, p = 0.55) or group × condition interaction (F = 0.43, p = 0.51).

Next, we estimated a sunk cost effect score for each participant based on their behavior choice. The sunk cost effect score was significantly lower in the ASD group than in the TD group (TD 0.30 ± 0.42, ASD 0.07 ± 0.28, p = 0.038) (Fig. 3). Considering that 11 participants with ASD took psychotropic drugs (see Supplementary Results for details), we compared the sunk cost effect score of participants with ASD who were not taking psychotropic drugs (n = 13) with that of the TD participants. The analy- sis did not materially change the result; the sunk cost effect score among participants with ASD who were not taking psychotropic drugs was significantly attenuated compared with the TD group (p = 0.030).

Thereafter, we performed correlation analyses between the sunk cost effect score and clinical symptom severity among the ASD participants. We found no significant rela- tionship between the sunk cost effect score and clinical symptom severity (all, p > 0.75).

Discussion

This study investigated decision-making under sunk costs in progress decision situations in individuals with ASD. Our results agree with the previous evidence of reduced sensitiv- ity to context stimuli in ASD and help us obtain a broader picture of the impact of sunk costs on their decision-making.

The TD group exhibited a sunk cost effect in the current task. Given that an invested sunk cost cannot be recovered, a rational forward-looking decision maker is expected to ignore sunk costs [18–22]. However, under the experimen- tal sunk cost condition in this study, the TD participants were more willing to make investment decisions than in the control condition.

Consistent with our hypothesis, the sunk cost effect was attenuated in the ASD group compared with the TD group. Notably, unlike the TD group, the rates of investments were not significantly increased in the sunk cost condition com- pared with the control condition. Individuals with ASD have repeatedly shown atypical performance on tasks that require processing of local information independently of its context [43, 44]. Many of the previous studies focused on percep- tual tasks, such as visual search, pitch discrimination, and motion-coherence detection, with corresponding theoreti- cal frameworks that highlight “low-level” processes, such as enhanced perceptual discrimination [13, 45]. Recently, studies on healthy subjects have demonstrated substantial overlap in brain activation between perceptual processing and “higher level” cognitive processing, such as decision- making and moral judgment, suggesting that social cognitive processing abilities depend at least partially on perceptual processing [46, 47]. Similarly, recent studies of ASD have reported that the reduced context sensitivity that character- izes ASD can be observed in “high-level” decision-making tasks. Specifically, the previous studies show that individuals with ASD are less susceptible to the framing and attraction effects [13, 31, 32]. As for the sunk cost effect, a recent study examined the effect in utilization decisions in individuals with ASD [30]. This study asked the participants to choose between two differently priced, but already-paid-for trips that coincidentally took place on the same day. Despite a stimulated preference for the cheaper trip, higher sunk costs for one of the alternatives significantly increased its con- sumption likelihood. This tendency was attenuated in par- ticipants with ASD [30]. Taken together, our results agree with these previous experimental findings and demonstrate that the reduced sunk cost effect, indicative of more consist- ent and conventionally rational choices, can be observed in progress as well as utilization decisions among individuals with ASD.

The current findings have practical implications on the economic and social functioning of individuals with ASD,

Fig. 3 Sunk cost effect score in the typical development (TD) and autism spectrum disorder (ASD) groups. The sunk cost effect score was significantly lower in the ASD group than in the TD group (p = 0.038). *p < 0.05

1069European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

because the sunk cost effect has an impact on many types of socioeconomic behavior. Although a heuristic approach, such as “past investments predict future benefits,” might help in many everyday decisions [21, 48], an overgener- alization could lead to severe financial or political con- sequences, such as the continuation of an unprofitable building project or even war [23–27]. The current results suggest that individuals with ASD might outperform those with TD in sunk-cost situations. To prevent serious con- sequences, it is important to elucidate individual differ- ences in the propensity to ignore sunk costs. Although some factors are reportedly associated with the sunk cost effect [24, 26], this issue remains largely unknown. Our findings might offer clues for studying the mechanisms generating the sunk cost effect, and for placing the suitable type of people in organizational jobs when ignoring sunk costs is beneficial.

Elucidating the heterogeneity of symptom expression in ASD is important for obtaining better understanding of the underlying neurobiological mechanisms and for the establishment of precise treatment strategies [14, 45, 49]. A behavioral economics approach can help elucidate existing symptomatology, or inform the development of new medi- ating markers and the personalization of treatment [5, 15, 50]. We did not identify any significant correlations between the strength of the sunk cost effect and severity of clini- cal symptoms among participants with ASD in this study. However, the distributions of the sunk cost effect score in our task in ASD participants were diverse. Comprehensive measurements with various tools of behavioral economics in larger samples would lead to a better understanding of the heterogeneity of altered decision-making in ASD.

Although there were mixed findings [51, 52], several previous studies reported that individuals with ASD chose risky options less often than control participants in the experimental gambling tasks [53, 54]. Unexpectedly, over- all rates of investments under the control condition in our task were higher in the ASD group than in the TD group. In most previous studies, the participants were instructed to choose between a sure and a gamble option in the gambling paradigms [52, 53]; conversely, the current task instructed participants to decide whether they wanted to invest the requested amount in the given project. The difference of the experimental paradigm formats may create fundamental dif- ferences in the strategies implemented by the participants. A recent study showed that ASD participants chose more rational options than control participants when risk aver- sion was the most rational strategy; however, both groups chose similarly when risk aversion was the less rational strategy [53]. The authors proposed that ASD participants relied more heavily on a risk-averse strategy, which may have motivated them to select more rational choices when the risk-averse choice was more rational; however, when the

risky choice was more rational, they overcame their risk- averse tendency [53]. Choosing the “Invest” option under the control condition was economically more rational in the current task, because the expected values of the “Invest” options were always higher than those of the “No invest” options for all six combinations of project costs and prob- abilities of success. Therefore, the ASD participants might frequently choose “Invest” options. This speculation should be examined in a variety of risky scenarios in the future.

This study has several limitations. First, nearly, half of the participants with ASD were administered psychotropic med- ication; thus, we cannot exclude the possibility of a medi- cation effect. For example, dopaminergic agents influence value-based and risky decision-making [55, 56]. Previous studies have shown that serotoninergic agents affect the dis- covery of bad decision outcomes and risk-seeking behavior [57]. Beninger et al. found an association between the status of antipsychotic drug use and better performance in gam- bling in patients with schizophrenia [58]. Unfortunately, the medicated participants with ASD in our study were admin- istered different types of psychotropic drugs, preventing any further analysis of the medication effects on behavioral results. However, the sunk cost effect score of participants with ASD who were not taking psychotropic drugs was also significantly reduced compared with that of the TD partici- pants. Second, we used a hypothetical choice experiment to understand actual choices under the influence of sunk costs. Although decision-making regarding hypothetical rewards does not necessarily reflect real-life decision-mak- ing behavior, the validity of the results of experiments with hypothetical rewards has been reported [59–61]. Therefore, we consider our findings to be useful in understanding real choices under the influence of sunk costs. Third, although individuals with ASD generally experience difficulties in the social environment [13, 14, 34], we used an economic task of a relatively nonsocial situation. Future studies should investigate whether the reduced sunk cost effect in ASD can be found in more social situations. A recent study conducted experiments involving interpersonal sunk cost scenarios as well as those involving intrapersonal ones [62]. The study has shown that people alter their choices in response to other people’s past investments [62]. In addition, the previ- ous studies investigated the influence of instruction on the sunk cost effect [26, 63, 64]. Extending such research to individuals with ASD could provide interesting information regarding the relationship between their social difficulties and sunk cost decision-making. Fourth, our ASD sample consisted of only high-functioning individuals with ASD. Future studies recruiting more ASD individuals with diverse IQ levels and not taking medication are required to replicate and strengthen our findings.

Notwithstanding these limitations, the current results extend the previous findings showing that the reduced

1070 European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

context sensitivity that characterizes ASD can be observed in sunk cost decision-making under progress decision situ- ations. Our findings contribute to a better understanding of ASD and may be useful in addressing practical implications on their socioeconomic behavior.

Acknowledgements The authors wish to extend their gratitude to the research team of the Medical Institute of Developmental Disabilities Research at Showa University for their assistance in data acquisition. This work was supported by grants-in-aid for scientific research A (24243061), C (17K10326), Young Scientists B (17K16398), and on Innovative Areas (23120009, 16H06572), from the Ministry of Educa- tion, Culture, Sports, Science and Technology of Japan (MEXT); and the Takeda Science Foundation. A part of this study is the result of the Strategic Research Program for Brain Sciences (JP19dm0107151) by Japan Agency for Medical Research and Development, “Research and development of technology for enhancing functional recovery of elderly and disabled people based on non-invasive brain imaging and robotic assistive devices,” the Commissioned Research of National Institute of Information and Communications Technology, JAPAN, and the Joint Usage/Research Program of Medical Institute of Devel- opmental Disabilities Research, Showa University. These agencies had no further role in the study design, the collection, analysis, and inter- pretation of data, the writing of the report, or in the decision to submit the paper for publication.

Author contributions JF, ST, TI, YYA, HO, R-IH, MN, NK, and HT designed research; JF, ST, TI, and YYA participated in the data acquisi- tion; JF, YYA, HO, MN, and NK were in charge of the clinical assess- ment. JF and ST analyzed data; TI, YYA, HO, MK, R-IH, MN, NK, and HT helped with interpretation of data. JF, ST, TI, YYA, HO, MK, R-IH, MN, NK, and HT wrote the paper. All authors have made intel- lectual contribution to the work and approved the final version of the manuscript for submission.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of interest.

Ethical standards The study was approved by the institutional ethical review board and has, therefore, been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Informed consent All participants gave their informed consent to par- ticipate prior to inclusion in the study.

References

1. Fond G, Bayard S, Capdevielle D et al (2013) A further evaluation of decision-making under risk and under ambiguity in schizophre- nia. Eur Arch Psychiatry Clin Neurosci 263:249–257

2. Fujino J, Tei S, Jankowski KF et al (2017) Role of spontaneous brain activity in explicit and implicit aspects of cognitive flex- ibility under socially conflicting situations: a resting-state fMRI study using fractional amplitude of low-frequency fluctuations. Neuroscience 367:60–71

3. Luk MSK, Chang WC, Chong CSY et al (2019) Altered risky decision making in patients with early non-affective psychosis. Eur Arch Psychiatry Clin Neurosci [Epub ahead of print]

4. Schmidt SNL, Fenske SC, Kirsch P et al (2019) Nucleus accum- bens activation is linked to salience in social decision making. Eur Arch Psychiatry Clin Neurosci 269:701–712

5. Sharp C, Monterosso J, Montague PR (2012) Neuroeconomics: a bridge for translational research. Biol Psychiatry 72:87–92

6. Tei S, Kauppi JP, Fujino J et al (2019) Inter-subject correlation of temporoparietal junction activity is associated with conflict pat- terns during flexible decision-making. Neurosci Res 144:67–70

7. Brosnan M, Lewton M, Ashwin C (2016) Reasoning on the autism spectrum: a dual process theory account. J Autism Dev Disord 46:2115–2125

8. Brosnan M, Ashwin C, Lewton M (2017) Brief report: intuitive and reflective reasoning in autism spectrum disorder. J Autism Dev Disord 47:2595–2601

9. Damiano CR, Aloi J, Treadway M et al (2012) Adults with autism spectrum disorders exhibit decreased sensitivity to reward param- eters when making effort-based decisions. J Neurodev Disord 4:13

10. Mussey JL, Travers BG, Klinger LG, Klinger MR (2015) Deci- sion-making skills in ASD: performance on the Iowa Gambling Task. Autism Res. 8:105–114

11. Vella L, Ring HA, Aitken MR et al (2018) Understanding self- reported difficulties in decision-making by people with autism spectrum disorders. Antism 22:549–559

12. Zhang L, Tang J, Dong Y et al (2015) Similarities and differences in decision-making impairments between autism spectrum disor- der and schizophrenia. Front Behav Neurosci 9:259

13. Farmer GD, Baron-Cohen S, Skylark WJ (2017) People with autism spectrum conditions make more consistent decisions. Psychol Sci 28:1067–1076

14. Fujino J, Tei S, Hashimoto RI et al (2017) Attitudes toward risk and ambiguity in patients with autism spectrum disorder. Mol Autism 8:45

15. Pushkarskaya H, Tolin D, Ruderman L et al (2015) Decision- making under uncertainty in obsessive–compulsive disorder. J Psychiatr Res 69:166–173

16. Takahashi H (2013) Molecular neuroimaging of emotional deci- sion-making. Neurosci Res 75:269–274

17. Weller RE, Avsar KB, Cox JE et al (2014) Delay discounting and task performance consistency in patients with schizophrenia. Psychiatry Res 215:286–293

18. Arkes HR, Blumer C (1985) The psychology of sunk cost. Organ Behav Hum Decis Process 35:124–140

19. Bogdanov M, Ruff CC, Schwabe L (2017) Transcranial stimula- tion over the dorsolateral prefrontal cortex increases the impact of past expenses on decision-making. Cereb Cortex 27:1094–1102

20. Friedman D, Pommerenke K, Lukose R et al (2007) Searching for the sunk cost fallacy. Exp Econ 10:79–104

21. Haller A, Schwabe L (2014) Sunk costs in the human brain. Neu- roimage 97:127–133

22. Thaler R (1980) Toward a positive theory of consumer choice. J Econ Behav Organ 1:39–60

23. Augenblick N (2016) The sunk-cost fallacy in penny auctions. Rev Econ Stud 83:58–86

24. Fujino J, Fujimoto S, Kodaka F et al (2016) Neural mechanisms and personality correlates of the sunk cost effect. Sci Rep 6:33171

25. Fujino J, Kawada R, Tsurumi K et al (2018) An fMRI study of decision-making under sunk costs in gambling disorder. Eur Neu- ropsychopharmacol 28:1371–1381

26. Roth S, Robbert T, Straus L (2015) On the sunk-cost effect in economic decision-making: a meta-analytic review. Bus Res 8:99–138

27. Zeng J, Zhang Q, Chen C et al (2013) An fMRI study on sunk cost effect. Brain Res 1519:63–70

28. Moon H (2001) Looking forward and looking back: integrating completion and sunk-cost effects within an escalation-of-commit- ment progress decision. J Appl Psychol 86:104–113

1071European Archives of Psychiatry and Clinical Neuroscience (2020) 270:1063–1071

1 3

29. Conlon DE, Garland H (1993) The role of project completion information in resource allocation decisions. Acad Manage J 36:402–413

30. Fujino J, Tei S, Itahashi T et al (2019) Sunk cost effect in individu- als with autism spectrum disorder. J Autism Dev Disord 49:1–10

31. De Martino B, Harrison NA, Knafo S et al (2008) Explaining enhanced logical consistency during decision making in autism. J Neurosci 28:10746–10750

32. Shah P, Catmur C, Bird G (2016) Emotional decision-making in autism spectrum disorder: the roles of interoception and alexithy- mia. Mol Autism 7:43

33. Luke L, Clare IC, Ring H et al (2012) Decision-making difficulties experienced by adults with autism spectrum conditions. Autism 16:612–621

34. Robic S, Sonie S, Fonlupt P et al (2015) Decision-making in a changing world: a study in autism spectrum disorders. J Autism Dev Disord 45:1603–1613

35. Fujino J, Tei S, Itahashi T et al (2019) Need for closure and cog- nitive flexibility in individuals with autism spectrum disorder: a preliminary study. Psychiatry Res 271:247–252

36. Tei S, Fujino J, Hashimoto RI et al (2018) Inflexible daily behav- iour is associated with the ability to control an automatic reaction in autism spectrum disorder. Sci Rep 8:8082

37. Tei S, Fujino J, Itahashi T et al (2019) Egocentric biases and atypi- cal generosity in autistic individuals. Autism Res [Epub ahead of print]

38. Lejuez C, Aklin WM, Jones HA et al (2003) The balloon analogue risk task (BART) differentiates smokers and nonsmokers. Exp Clin Psychopharmacol 11:26–33

39. Lord C, Risi S, Lambrecht L et al (2003) The autism diagnostic observation schedule—generic: a standard measure of social and communication deficits associated with the spectrum of autism. J Autism Dev Disord 30:205–223

40. Matsuoka K, Kim Y (2006) Japanese adult reading test (JART). Shinkou-Igaku Publishers, Tokyo

41. Matsuoka K, Uno M, Kasai K et al (2006) Estimation of premor- bid IQ in individuals with Alzheimer’s disease using Japanese ideographic script (Kanji) compound words: Japanese version of national adult reading test. Psychiatry Clin Neurosci 60:332–339

42. Tanaka Y, Fujino J, Ideno T et al (2014) Are ambiguity aversion and ambiguity intolerance identical?: a neuroeconomics investiga- tion. Front Psychol 5:1550

43. Jolliffe T, Baron-Cohen S (1997) Are people with autism and Asperger syndrome faster than normal on the embedded figures test? J Child Psychol Psychiatry 38:527–534

44. Plaisted K, O’Riordan M, Baron-Cohen S (1998) Enhanced visual search for a conjunctive target in autism: a research note. J Child Psychol Psychiatry 39:777–783

45. Happé F, Frith U (2006) The weak coherence account: detail- focused cognitive style in autism spectrum disorders. J Autism Dev Disord 36:5–25

46. Decety J, Lamm C (2007) The role of the right temporoparietal junction in social interaction: how low-level computational pro- cesses contribute to meta-cognition. Neuroscientist 13:580–593

47. Tei S, Fujino J, Kawada R et al (2017) Collaborative roles of tem- poroparietal junction and dorsolateral prefrontal cortex in different types of behavioural flexibility. Sci Rep 7:6415

48. Gigerenzer G, Goldstein DG et  al (1996) Reasoning the fast and frugal way: models of bounded rationality. Psychol Rev 103:650–669

49. Goddard MN, Swaab H, Rombouts SA et al (2016) Neural systems for social cognition: gray matter volume abnormalities in boys at high genetic risk of autism symptoms, and a comparison with idiopathic autism spectrum disorder. Eur Arch Psychiatry Clin Neurosci 266:523–531

50. Fujino J, Hirose K, Tei S et  al (2016) Ambiguity aversion in schizophrenia: an fMRI study of decision-making under risk and ambiguity. Schizophr Res 178:94–101

51. South M, Dana J, White SE et al (2011) Failure is not an option: risk-taking is moderated by anxiety and also by cognitive ability in children and adolescents diagnosed with an autism spectrum disorder. J Autism Dev Disord 41:55–65

52. Wu HC, White S, Rees G et al (2018) Executive function in high- functioning autism: decision-making consistency as a character- istic gambling behaviour. Cortex 107:21–36

53. Gosling CJ, Moutier S (2018) Brief Report: risk-aversion and rationality in autism spectrum disorders. J Autism Dev Disord 48:3623–3628

54. Levin IP, Gaeth GJ, Foley-Nicpon M et al (2015) Extending deci- sion making competence to special populations: a pilot study of persons on the autism spectrum. Front Psychol 6:539

55. Doya K (2008) Modulators of decision making. Nat Neurosci 11:410–416

56. St Onge JR, Floresco SB (2009) Dopaminergic modulation of risk-based decision making. Neuropsychopharmacol 34:681–697

57. Rogers RD (2011) The roles of dopamine and serotonin in deci- sion making: evidence from pharmacological experiments in humans. Neuropsychopharmacol 36:114–132

58. Beninger RJ, Wasserman J, Zanibbi K et al (2003) Typical and atypical antipsychotic medications differentially affect two non- declarative memory tasks in schizophrenic patients: a double dis- sociation. Schizophr Res 61:281–292

59. Bray S, Shimojo S, O’Doherty JP (2010) Human medial orbito- frontal cortex is recruited during experience of imagined and real rewards. J Neurophysiol 103:2506–2512

60. Kang MJ, Rangel A, Camus M (2011) Hypothetical and real choice differentially activate common valuation areas. J Neurosci 31:461–468

61. Locey ML, Jones BA, Rachlin H (2011) Real and hypothetical rewards. Judgm Decis Mak 6:552–564

62. Olivola CY (2018) The interpersonal sunk-cost effect. Psychol Sci 29:1072–1083

63. Hafenbrack AC, Kinias Z, Barsade SG (2014) Debiasing the mind through meditation: mindfulness and the sunk-cost bias. Psychol Sci 25:369–376

64. Tan HT, Yates JF (1995) Sunk cost effects: the influences of instruction and future return estimates. Organ Behav Hum Decis Process 63:311–319

European Archives of Psychiatry & Clinical Neuroscience is a copyright of Springer, 2020. All Rights Reserved.