A multi-pronged investigation of option generation using depression, PET and modafinil

Abstract

Option generation is a critical process in decision making, but previous studies have largely focused on choices between options given by a researcher. Consequently, how we self-generate options for behaviour remain poorly understood. Here, we investigated option generation in major depressive disorder and how dopamine might modulate this process, as well as the effects of modafinil (a putative cognitive enhancer) on option generation in healthy individuals. We first compared differences in self-generated options between healthy non-depressed adults [n = 44, age = 26.3 years (SD 5.9)] and patients with major depressive disorder [n = 54, age = 24.8 years (SD 7.4)]. In the second study, a subset of depressed individuals [n = 22, age = 25.6 years (SD 7.8)] underwent PET scans with 11C-raclopride to examine the relationships between dopamine D2/D3 receptor availability and individual differences in option generation. Finally, a randomized, double-blind, placebo-controlled, three-way crossover study of modafinil (100 mg and 200 mg), was conducted in an independent sample of healthy people [n = 19, age = 23.2 years (SD 4.8)] to compare option generation under different doses of this drug. The first study revealed that patients with major depressive disorder produced significantly fewer options [t(96) = 2.68, P = 0.009, Cohen's d = 0.54], albeit with greater uniqueness [t(96) = -2.54, P = 0.01, Cohen's d = 0.52], on the option generation task compared to healthy controls. In the second study, we found that 11C-raclopride binding potential in the putamen was negatively correlated with fluency (r = -0.69, P = 0.001) but positively associated with uniqueness (r = 0.59, P = 0.007). Hence, depressed individuals with higher densities of unoccupied putamen D2/D3 receptors in the putamen generated fewer but more unique options, whereas patients with lower D2/D3 receptor availability were likely to produce a larger number of similar options. Finally, healthy participants were less unique [F(2,36) = 3.32, P = 0.048, partial η2 = 0.16] and diverse [F(2,36) = 4.31, P = 0.021, partial η2 = 0.19] after taking 200 mg versus 100 mg and 0 mg of modafinil, while fluency increased linearly with dosage at a trend level [F(1,18) = 4.11, P = 0.058, partial η2 = 0.19]. Our results show, for the first time, that option generation is affected in clinical depression and that dopaminergic activity in the putamen of patients with major depressive disorder may play a key role in the self-generation of options. Modafinil was also found to influence option generation in healthy people by reducing the creativity of options produced.

Keywords: depression; dopamine; modafinil; option generation; raclopride PET.

Figure 1

Paradigms and quantitfication of uniqueness and diversity. (A) Option generation task. Participants were given 1.5 min to draw as many and as different paths as they could between two fixed red circles on a touchscreen computer. (B) Quantifying uniqueness and diversity. Each path was divided into 200 points equally spaced along its length in order to derive a feature vector that comprised the position, first derivatives (which accounted for slopes) and second derivatives (which accounted for curvatures). The ‘distance’ between any two paths was computed by subtracting the features of one path from the other; and the uniqueness of each path generated by every participant was then taken to be the ‘distance’ between it and the most similar path produced by all other subjects in the three studies of this paper. Multi-dimensional scaling was also used to project the pairwise distance matrix of every participant onto a 2D subspace, and diversity was approximated by the area of the convex hull covering these points. (C) Motor execution control task. To assess baseline drawing speed, participants were asked to produce 10 straight lines, each as quickly as they could, between the two fixed circles. (D) Externally-cued action control task. To account for motor planning, participants were required to draw a straight line from the bottom red circle to a random target location decided by the computer. A new target location was presented after the completion of each path, and subjects had to connect to as many target locations as possible in 45 s. (E) Option selection control task. To assess option selection ability, participants were required to choose an option from a set of displayed target locations and then draw a straight line from a central start location to it. The goal was to make as many connections as possible in 45 s. Figure modified from Ang et al.

Figure 2

Comparison of (A) fluency, (B) uniqueness and (C) diversity between healthy controls (HC) and patients with MDD. After accounting for performance on three controls tasks, the patients with MDD were found to have generated significantly fewer options compared with the HCs. However, they exhibited greater uniqueness in the paths produced, suggesting that the depressed patients were biased toward generating fewer options but with higher mean uniqueness. There was no difference in diversity, indicating that the options produced by both groups were similarly varied. **P < 0.01, *P < 0.05.

Figure 3

Correlation between putamen D₂/D₃ receptor availability and (A) fluency, (B) uniqueness as well as (C) diversity in patients with MDD. Putamen D₂/D₃ receptor availability, as indexed by ¹¹C-raclopride BP_ND, was negatively associated with fluency but positively correlated with uniqueness. There was also a positive trending relationship between putamen D₂/D₃ receptor availability and diversity. Note that age and gender have been regressed out from BP_ND. Performance on the control tasks have also been partialled out from the option generation metrics.

Figure 4

Comparison of (A) fluency, (B) uniqueness and (C) diversity between healthy volunteers on placebo, 100 mg and 200 mg of modafinil. After controlling for performance on the control tasks, there was no significant effect of dose on fluency. However, participants generated less unique and less varied options after taking 200 mg of modafinil compared with placebo and 100 mg of modafinil. **P < 0.01, *P < 0.05, ^†P < 0.10.