The task consisted of 20 discrete choices between a smaller immediate reward presented in a box on the left side of the screen and an 80¢ delay reward presented in a box on the right side of the screen . At the top center of the screen was a box displaying total earnings on the task. On any trial, if the smaller sooner reward was selected with a single mouse click, the response options disappeared and a button appeared that stated “Click here to bank your amount.” Upon a single mouse click on this button, that amount was dispensed from the coin dispenser, and the total earnings box was updated. If the delayed 80¢ was selected, the response options disappeared and a number in the middle of screen counted down the number of seconds to wait . When the delay elapsed, a button appeared that required the participant to click to “bank” the 80 cents, at which point the coins were delivered and the total earnings were updated. When money was delivered, participants removed the coins from the dispensing tray and dropped them into a glass jar. There were 5 blocks of 4 trials each, with each block associated with a different delay for the 80 cent reward. The delays were 5, 10, 20, 40, and 80 s, and followed an increasing order across blocks. On the first trial of each block, the immediate reward size was 40¢ . The smaller reward was then adjusted within the block using a “decreasing adjustment” algorithm, which has been used in previous human studies involving hypothetical rewards . Specifically, the smaller sooner reward was adjusted by 20, 10, and 5¢ on Trials 2, 3, and 4 of the block, respectively, in the direction that would move choice toward indifference . The indifference point was defined as the value that would have been presented on a 5th trial had the algorithm continued . Indifference points therefore varied by increments of 2.5¢,vertical grow rack and were divided by 80¢ to be expressed as the proportion of the larger reinforcer. Indifference points were expressed as a proportion of the larger later reward .
A waiting period was imposed after the final trial to prevent participants from choosing the smaller immediate reward to end the task or session sooner . Participants were told before beginning the task that the total duration of the task would be independent of the choices made during the task, although participants were not explicitly told about the waiting period at the end of the task that was responsible for ensuring approximately equal task duration. The waiting period was defined as 660 s minus the sum of all larger reward delays that the participant experienced throughout the task. Although this manipulation ensured that total programmed waiting time did not substantially differ across participants, differences in participant response latency nonetheless allowed for some variability in total task time. At the end of the task, participants exchanged whole dollar amounts of coins for paper currency.The schizophrenia and control groups had qualitatively similar DD functions, but quantitatively, the schizophrenia group showed a significantly greater DD than controls on the experiential task, and normal DD on the hypothetical task. The schizophrenia group’s performance on the DD tasks was generally not associated with a range of potential confounds. In addition, test-retest reliability was examined for the schizophrenia group and was good on both tasks. These findings provide the first evidence of impaired DD in schizophrenia using an experiential paradigm that parallels tasks used in animal research much more closely than conventional human paradigms. While not all aspects of reward processing are impaired in schizophrenia , these findings suggest alterations do extend to a delay discounting context that involves real rewards and real delay periods. As described below, the schizophrenia group’s pattern of altered experiential and normal hypothetical DD likely reflects the fact the these tasks differed on several key dimensions, including reward type , reward magnitude , and delay time frame . Regarding qualitative analyses, the shape and orderliness of the DD data were generally similar across groups. In line with a prior report , the schizophrenia group showed typical hyperbolic discounting functions across tasks. Further, a large majority demonstrated orderly data for both DD tasks. The proportion with less-orderly data on the hypothetical, though not the experiential, task was significantly larger than controls.
However, the main study findings were unchanged after removing the subset of participants from both groups with less-orderly data. In this first study of experiential DD in schizophrenia, the schizophrenia group showed quantitatively greater discounting than controls for actual monetary rewards delivered in real time. Diminished discounting on this and similar experiential tasks has been reported in other clinical populations, including cocaine dependence, ADHD, and smokers . Experiential tasks appear to tap into a rather different aspect of DD than hypothetical tasks. For example, the correlation between hypothetical and experiential DD tasks was relatively small in both groups. Several studies have also reported relatively low convergence between these tasks and one found altered experiential but not hypothetical discounting in ADHD . There were no quantitative group differences for the hypothetical DD task and this study included the largest schizophrenia and control samples examined to date. Our finding on this task is consistent with two prior studies , but inconsistent with four others that found greater hypothetical DD in schizophrenia . The rather substantial methodological differences across the few DD studies make it difficult to pinpoint why three studies found normal DD but four did not. Since all prior studies included chronically ill samples, and all except one examined outpatients, the discrepancies across studies are not attributable to these participant characteristics. However, the tasks and data analytic approaches varied widely. For example, across the seven studies, the maximum delayed reward magnitude ranged from $86 to $1000, and the maximum delayed reward duration ranged from a few months up to 50 years. Further research will want to systematically assess the impact of these parameters on hypothetical DD in schizophrenia. For example, it could be informative to examine how individuals with schizophrenia perform on a hypothetical task with reward magnitudes and delay intervals that correspond to those in the experiential task. The current study considered a wide range of potentially confounding factors on DD and found that their impact was small. The only relevant factor was smoking status.
Smokers showed greater hypothetical DD than non-smokers, which converges with prior findings from the general population and schizophrenia . However, we still found the pattern of altered experiential and normal hypothetical DD in schizophrenia when we limited our analyses to non-smokers. There were no significant associations between DD and other substances, symptoms, or anti-psychotic medication dosages. Given the conceptual link between reward processing and negative symptoms , it is somewhat puzzling that alterations in DD, particularly on the experiential task, did not significantly correlate with higher clinically rated negative symptoms. Although some studies have found that neuroscience-based reward and decision making tasks are associated with negative symptoms a number of studies by our group and others failed to detect such relationships . The reason for these discrepancies is not year clear. We have suggested that there are complex intervening steps on the causal pathway between the relatively discrete processes measured by decision-making tasks and the broad aspects of experience and behavior that are captured by clinical rating scales,cannabis grow racks which may substantially diminish direct correlations . DD also showed no significant associations with global or particular domains of neurocognition. This does not support prior suggestions that DD disturbances in schizophrenia reflect problems in the representation and maintenance of reward value . The schizophrenia group’s pattern of altered experiential but normal hypothetical DD was also not attributable to differences in the test-retest reliabilities of the tasks. The test-retest correlations of approximately .70 for both tasks are similar to prior reports in healthy samples and the group means showed good stability across occasions. These findings, in conjunction with the lack of associations with symptoms, suggest the DD tasks are measuring relatively stable traits among individuals with schizophrenia. These properties support the use of the experiential DD task as a performance measure of decision-making impairment in clinical trials for schizophrenia . Its potential usefulness for clinical trials is bolstered by evidence that it is sensitive to state-related changes, such as sleep deprivation, dopamine agonist administration in Parkinson’s disease, alcohol administration, and methylphenidate administration in ADHD . One might have expected the schizophrenia group to show greater difficulties for hypothetical, distant rewards in light impaired abstract thinking and longer-term prospection associated with this disorder . However, the pattern found in the current study may relate to participant and task characteristics. Regarding participant characteristics, since schizophrenia is associated with decreased SES and many in the schizophrenia group were receiving limited fixed incomes, the schizophrenia group may have valued immediately available, real rewards more than controls. This possibility is bolstered by our finding that the schizophrenia group assigned higher value ratings than controls for the lowest value but similar ratings for the highest value on the subjective valuation of money index, and with previous research showing greater discounting in lower income adults . Although individual differences in subjective valuation ratings did not significantly correlate with performance on the DD tasks, this factor remains a possible contributor . Regarding task characteristics, Paglieri postulated key differences between hypothetical tasks and experiential tasks, beyond reward magnitude and delay length.
Whereas hypothetical tasks merely involve postponing receipt of a reward with no constraints on how subjects spend their time during the intervening delay, the waiting period in experiential tasks comes with associated costs. These include direct costs, such as boredom or discomfort, and opportunity costs, such as valuable activities that the participant could be engaged in if not forced to wait. The relevance of such costs was demonstrated in a recent study that found DD rates increased as an orderly function of the constraints on what people could do during the delay interval on a hypothetical task . Perhaps the individuals with schizophrenia in our study were hyper-responsive to the associated costs of doing nothing in the delay period and experienced alterations in their cost/benefit calculations. For example, schizophrenia is associated with an elevated tendency to experience negative affect/arousal and boredom , as well as altered decision-making on tasks that involve weighing the relative effort expenditure costs against monetary rewards . Studies that manipulate the constraints, or obtain subjective ratings/ psychophysiological measures, during delay intervals could shed light on the possible impact of these costs in DD in schizophrenia. Strengths of the current study include the large clinical sample, use of two different types of DD tasks, rigorous evaluation of data integrity, examination of many potential confounds, and evaluation of test-retest reliability. However, the study has some limitations and highlights areas in need of further study. First, participants with schizophrenia were taking medications at clinical dosages. Although dosage equivalents were not related to DD, the impact of medications remains unclear. Second, the schizophrenia sample was chronically ill and it is unknown whether similar DD patterns would be evident in younger or high-risk samples. Third, the order of delay discounting task administration was not counterbalanced, so we are unable to examine potential order effects. Fourth, although performance on the tasks was not related to subjective valuation of money, we did not obtain objective measures to evaluate whether income or socio-economic status was associated with DD task performance. Fifth, this study only assessed monetary rewards and it is unknown whether similar patterns would be found for other primary or secondary reinforcers. Sixth, although the schizophrenia group showed normal performance on the hypothetical DD task, we cannot tell if the normal choice patterns were achieved through abnormal neural processes. For example, a small fMRI study reported that individuals with schizophrenia showed an abnormal hypo-activation in some regions and hyper-activation in others while making DD decisions . Further attention to these issues can help clarify the nature of impaired reward processing and decision-making in schizophrenia. General Scientific Summary: Delay discounting refers to whether one is willing to forego a smaller, sooner reward for the sake of a larger, later reward. This study found that people with schizophrenia showed a greater preference for smaller, sooner rewards than healthy comparison participants on a DD task that involved making choices about actual monetary rewards provided in real time.