Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Apr;4(3):193-202.
doi: 10.1089/brain.2013.0184. Epub 2014 Apr 4.

Age-related Differences in Advantageous Decision Making Are Associated With Distinct Differences in Functional Community Structure

Affiliations
Free PMC article

Age-related Differences in Advantageous Decision Making Are Associated With Distinct Differences in Functional Community Structure

Malaak Nasser Moussa et al. Brain Connect. .
Free PMC article

Abstract

Human decision making is dependent on not only the function of several brain regions but also their synergistic interaction. The specific function of brain areas within the ventromedial prefrontal cortex has long been studied in an effort to understand choice evaluation and decision making. These data specifically focus on whole-brain functional interconnectivity using the principles of network science. The Iowa Gambling Task (IGT) was the first neuropsychological task used to model real-life decisions in a way that factors reward, punishment, and uncertainty. Clinically, it has been used to detect decision-making impairments characteristic of patients with prefrontal cortex lesions. Here, we used performance on repeated blocks of the IGT as a behavioral measure of advantageous and disadvantageous decision making in young and mature adults. Both adult groups performed poorly by predominately making disadvantageous selections in the beginning stages of the task. In later phases of the task, young adults shifted to more advantageous selections and outperformed mature adults. Modularity analysis revealed stark underlying differences in visual, sensorimotor and medial prefrontal cortex community structure. In addition, changes in orbitofrontal cortex connectivity predicted behavioral deficits in IGT performance. Contrasts were driven by a difference in age but may also prove relevant to neuropsychiatric disorders associated with poor decision making, including the vulnerability to alcohol and/or drug addiction.

Figures

<b>FIG. 1.</b>
FIG. 1.
Iowa Gambling Task (IGT). Participants were presented with 4 decks (A, B, C, and D) during card selection, and given monetary win and loss information during feedback. Disadvantageous decks A and B resulted in immediate gains of $100 with losses over time ranging from $250 to $1200. Advantageous decks C and D produced immediate gains of $50 with losses over time ranging from $50 to $250.
<b>FIG. 2.</b>
FIG. 2.
IGT performance across trial. Both adult groups performed poorly in Block 1 of the IGT. In Block 3, however, young adults did significantly better than mature adults. There was a main effect of adult group [GLM, F (1, 16)=9.3, p=0.01], no main effect of block and a significant block × adult group interaction [GLM, F (1, 16)=5.8, p=0.03]. Block Net Score was calculated as advantageous (C+D)—disadvantageous (A+B) deck selections. GLM, general linear model.
<b>FIG. 3.</b>
FIG. 3.
Differences in card selection that are based on overall card value. Spearman Rank correlation coefficients (ρ) were determined using one lag autocorrelations; decks A, B, C, and D represented the abscissa and the overall value (i.e., wins − losses) of the previous card represented the ordinates. Individual subject ρ values were calculated and averaged across young and mature adults for Blocks 1 and 3. A negative ρ indicated that disadvantageous selections were made after cards selections with high overall value. A positive ρ indicated that advantageous selections were made after cards selections with high overall value. In Block 1, young and mature adults picked from more disadvantageous decks immediately after card selections with high overall value. In Block 3, the value of the preceding card no longer related to subsequent selection of either advantageous or disadvantageous card types in young adults. Mature adults, on the other hand, continued to pick more disadvantageous decks immediately after card selections with high overall value. The difference in ρ across blocks was used to measure this change in behavior, and was significantly different between young and mature adults [two-tailed t-test, t (16)=3.27, p=0.01].
<b>FIG. 4.</b>
FIG. 4.
Representative young and mature adult selections. Representative deck selections (black circle), card wins (green triangle), and card losses (red square) are shown for each event in a block for (A) one young adult and (B) one mature adult. In Block 1, both adult groups start off by sampling decks. In Block 3, young adults picked advantageous decks (C and D) and occasionally made a disadvantageous selection. Mature adults selected disadvantageous decks almost exclusively in Block 3, and in particular preferred Deck B. Shortly following card selections with high loss (highlighted red squares) mature adults immediately chose advantageous decks. Shortly thereafter, however, mature adults reverted to selecting Deck B. Color images available online at www.liebertpub.com/brain
<b>FIG. 5.</b>
FIG. 5.
Visual and sensorimotor functional community structure. Visual community consistency was high in both blocks for young adults. In mature adults, visual community consistency was higher in the last block relative to the first. Sensorimotor community consistency in both young and mature adults was higher in the last block relative to the first. SI, scaled inclusivity. Color images available online at www.liebertpub.com/brain
<b>FIG. 6.</b>
FIG. 6.
Medial prefrontal cortex (mPFC) functional community structure. In Block 1, the young adult mPFC community was confined. In mature adults, it was distributed and included the precuneus and posterior cingulate cortices. In Block 3, the mPFC community in young adults was no longer confined and like mature adults included the precuneus, posterior cingulate, and parietal cortices. Color images available online at www.liebertpub.com/brain
<b>FIG. 7.</b>
FIG. 7.
Orbitofrontal cortex (OFC) functional community structure. OFC community consistency was only detected in young adults during successful IGT performance in Block 3. Color images available online at www.liebertpub.com/brain
<b>FIG. 8.</b>
FIG. 8.
Change in connectivity in the OFC and relation to performance. (A) At each node in the network, the number of new and lost links across blocks was calculated. A net loss of links within the OFC was found in mature adults. Young adults showed a slight positive net change [mean new-lost OFC links young adults±standard error: 5342±18630 and mean new-lost OFC links mature adults: −23639±13021; two-tailed t-test, t (16)=1.8, p=0.22]. (B) New-lost OFC links showed a positive correlation with Net Score Difference (Pearson's R=0.41, one-tailed, p=0.04). Color images available online at www.liebertpub.com/brain

Similar articles

See all similar articles

Cited by 5 articles

Publication types

Feedback