Decision-Making Under Ambiguity and Risk in Adolescents and Young Adults with Autism Spectrum Disorder

Friday, May 12, 2017: 1:57 PM
Yerba Buena 8 (Marriott Marquis Hotel)
M. K. Krug1, C. C. Coleman2, G. C. Gower2 and M. Solomon3, (1)MIND Institute, Sacramento, CA, (2)Department of Psychiatry & Behavioral Sciences, University of California-Davis, Sacramento, CA, (3)Department of Psychiatry & Behavioral Sciences, MIND Institute, Sacramento, CA

Little is known about decision-making in autism spectrum disorder (ASD), despite its potential impact on functioning. Decision-making under ambiguity can be investigated using the Iowa Gambling Task (IGT), where learning is implicit. Studies have shown impaired performance on the IGT in ASD (Mussey et al., 2015; Zhang et al., 2015), although South et al. (2014) observed enhanced performance. A modified version of the IGT (mIGT), which eliminates the potential confound of differences in exploratory behavior, and allows for separation of approach (choosing to “play” a good deck) and avoidance (choosing to “pass” on a bad deck) behavior, has been developed but has not yet been used in ASD (Cauffman et al., 2010). Decision-making under risk can be assessed using the Game of Dice Task (GDT), where outcomes are unambiguous and explicit. Zhang et al. (2015) found impaired performance and suboptimal use of trial feedback in ASD.


1. Continue to investigate decision-making under risk using the GDT. 2. Examine decision-making under ambiguity using the new mIGT. 3. Explore how performance on decision-making tasks is related to cognitive functioning and psychopathology.


22 ASD (Age = 16.45(2.67); FSIQ = 101.27(15.11)) and 29 TYP participants (Age = 16.07(2.91); FSIQ = 110.03(12.79)) performed the GDT and mIGT. In the GDT a bet is placed on 1, 2, 3 or 4 sides of the die. A selection of 1 or 2 sides is considered “risky,” while selection of 3 or 4 sides is considered “safe.” For each mIGT trial one of four decks is pre-selected and can be “played” or “passed.” Cognition was assessed using the NIH Toolbox Cognition Battery (www.nihtoolbox.org). Psychopathology was assessed with the Achenbach System of Empirically Based Assessment (ASEBA) (Achenbach & Rescorla, 2001; 2003).


GDT: Mean net score (#safe choices – #risky choices) was significantly higher for TYP (12.59(6.67)) versus ASD (7.27(8.56)), F (1,48) = 4.820, p = .033. TYP was more likely to place a safe bet following a “safe” win trial (F (1,48) = 5.275, p = .026) and after a “risky” loss trial (F (1,36) = 3.688, p = .063, trend) compared to ASD (Figure 1a), indicating better use of feedback.

 mIGT: For TYP, the difference in %played between good decks and bad decks was significant for Blocks 3, 5, and 6 (all ps <.05). A comparison of Block 1 and Block 6 indicated that TYP, but not ASD, participants learned to avoid bad decks (t(28) = 2.709, p = .011) (Figure 1b).

TYP performance on IGT was positively correlated with NIH Toolbox Cognitive Function Composite Score (Figure 2a). In ASD, there was an association between anxiety and obsessive compulsive symptoms and poorer IGT performance (Figure 2b, c).


Participants with ASD show impairments in decision making under risk (GDT) and ambiguity (mIGT). Decision making is related to cognitive functioning in TYP and anxiety and compulsive symptoms in ASD. Data collection is ongoing, and age groups effects analyses will assess development of decision-making processes throughout adolescence and young adulthood and their relationship to adaptive functioning.