H567/02 Section A - Casey et al. (2011): delay of gratification

Background, design and follow-up sample

Delay of gratification means resisting an immediate reward in favour of a later reward. Earlier preschool research compared taking a small treat immediately with waiting for a larger one. A hot/cool account proposed interacting processes: emotionally attractive “hot” cues encourage an immediate response, while “cool” cognitive control helps maintain a longer-term goal. The adult study asked whether consistently high and low delayers differed in impulse control, especially for attractive social cues, about 40 years later.

The longitudinal source cohort contained 562 children with preschool delay scores from Stanford's Bing Nursery School. Later self-control questionnaires were available from 155 in their twenties and 135 in their thirties; these are follow-up counts, not the adult task sample. Researchers identified 117 people whose preschool and available adult scores were consistently above or below the relevant averages and invited them. 59 accepted: 32 high delayers and 27 low delayers, with 36 women and 23 men overall, now approximately in their mid-forties.

Of those 59, 27 volunteered for imaging. One low-delaying man was excluded for poor task performance, leaving 26 in the imaging analyses: 15 high and 11 low delayers. Do not confuse the 27 who agreed to scanning with the 26 analysed.

The design combined longitudinal follow-up with a quasi-experimental comparison of pre-existing high/low groups. People were not randomly assigned to decades of high or low self-control. Within the adult task, cue and response conditions were varied, giving repeated-measures comparisons. The dependent measures were reaction time, correct responses/false alarms, and task-related fMRI signals. The study investigated impulse control through a new adult task; it did not repeat the preschool food-reward procedure in the scanner.

Go/nogo materials and procedure

In a go/nogo task, participants press a button to a designated target but withhold the response to a nontarget. Frequent go trials establish a tendency to respond, so successful withholding on a rarer nogo trial requires inhibition. A false alarm is an incorrect press on a nogo trial, not a slow correct press.

Participants received laptops at home, with standardised face stimuli from the NimStim set and computer instructions to respond quickly and accurately. A face appeared for 500 ms, followed by a one-second interval. Each run had 160 trials: 120 go and 40 nogo, presented in pseudorandom order. Targets were reversed across runs so the conditions were not permanently tied to one sex or expression.

Both expressions were used as targets and nontargets. “Hot” does not mean every happy face always required pressing, or that fear and happiness produced the same group difference.

In experiment 2 participants performed the hot version only during fMRI. Faces again appeared for 500 ms, but intervals varied from 2 to 14.5 seconds, averaging 5.2 seconds, to help separate responses in the imaging analysis. Each scanner run contained 48 trials: 35 go and 13 nogo. A rear-projection screen displayed the task and a five-button response pad recorded responses. Scanner timing and trial counts therefore differed from the home task.

Functional MRI measures changes in blood-oxygenation-related signals associated with neural activity. It is an indirect indicator of regional recruitment, not a direct recording of each neuron, a photograph of a thought or a structural measure of regional size.

Behavioural and imaging findings

Behavioural accuracy on go trials was high overall: 99.8% for cool and 99.5% for hot trials. High and low delayers did not differ significantly in reaction times or go accuracy. The informative contrast concerned failures to withhold a response.

Thus the findings do not mean that low delayers failed every inhibition task. In the scanner, false alarms again were higher in low delayers (14.5%) than high delayers (10.9%), but that behavioural group difference was not significant (p = .29). A similar direction is not the same as a statistically established difference.

Imaging associated the right inferior frontal gyrus, a prefrontal region, with correctly withholding a response compared with making one. A follow-up regional analysis found greater differentiation between nogo and go signals in high delayers than low delayers. This was not a significant delay-group effect everywhere in the whole-brain analysis.

The ventral striatum, implicated in reward-related processing, showed an interaction involving delay group, trial type and emotion. Low delayers showed greater recruitment in runs where happy faces were the nogo stimulus. The specific happy-nogo follow-up contrast was a trend (p = .06), although the interaction and overall happy-nogo-run contrast provided significant evidence. Keep the result attached to its comparison; do not declare every striatal group contrast significant.

Conclusions and brain networks

The authors concluded that resistance to temptation showed some long-term stability in this selected group, but that its behavioural expression depended on the cues to be resisted. Frontal control-related and striatal reward-related processes were both relevant. Reduced differentiation of inferior-frontal responses and stronger striatal recruitment in low delayers were interpreted as neural correlates of more difficulty resisting attractive cues.