Elsevier

Brain Research

Volume 1141, 13 April 2007, Pages 168-177
Brain Research

Research Report
Reward modulation of prefrontal and visual association cortex during an incentive working memory task

https://doi.org/10.1016/j.brainres.2007.01.052Get rights and content

Abstract

Cognitive performance differs with motivation, but little direct evidence exists regarding the neural mechanisms of the influence of reward motivation on working memory (WM). We tested the effects of motivation on the top-down control in visual WM. Encoding relevant stimuli for maintenance, while suppressing inappropriate inputs is considered a core process in cognition. Prior functional magnetic resonance imaging (fMRI) results demonstrated that stimulus-specific visual association cortex serves as a marker of activation differences for task-relevant and task-irrelevant inputs, such that enhanced activity occurs when attention is directed to relevant stimuli and suppressed activity occurs when attention is directed away from irrelevant stimuli [Gazzaley, A., Cooney, J., McEvoy, K., Knight, R.T., and D'Esposito, M. J. Cogn. Neurosci. 17, 507–517]. We used fMRI to test whether differential WM performance, indexed by lowered response times on a delayed-recognition task, was associated with amplification of enhancement and suppression effects during stimulus encoding within visual association cortex. Our results indicate that enhancement and suppression are amplified for trials with the highest reward level relative to non-rewarded trials for a scene-selective cortical region. In a face-selective region, similar modulation of enhancement for the highest reward level relative to non-rewarded trials was found. Prefrontal cortex also showed enhanced activity during high reward trials. Overall these results reveal that reward motivation can play a pivotal role in driving performance through top-down signaling in frontal regions involved in WM, as well as visual association regions selective to processing the perceptual inputs of the items to be remembered.

Introduction

There is a rich literature linking reward motivation and working memory (WM); however; it has largely consisted of single-cell recording in non-human primates. This evidence suggests that reward motivation and WM processes overlap and interact. Despite this linkage, there has been little direct investigation of the interaction between reward motivation and WM in humans.

Much of the research aimed at understanding the neural basis of reward processing has focused on the responsitivity of the midbrain dopaminergic system and limbic cortex in the anticipation and receipt of reward. For example, numerous studies have emphasized the involvement of both the nucleus accumbens (Breiter et al., 2001, Knutson et al., 2001) and ventral striatum (Elliott et al., 2000, Delgado et al., 2003) in reward processing. Additionally, the orbital sector of the prefrontal cortex (PFC) has been emphasized in representing reward value (O'Doherty et al., 2001, Delgado et al., 2003, Rolls, 2004, Izquierdo et al., 2004, Simmons et al., 2005). Several recent studies have also begun to probe the influence of cognitive aspects and how they modulate reward-related brain regions. Tricomi et al. (2004) showed that fMRI response in the caudate nucleus was dependent upon the need to perform an action to obtain reward. Zink et al. (2004) have presented similar evidence showing that modulation of the caudate and nucleus accumbens is dependent upon task-related action contingency when compared to the passive receipt of money.

Further investigations of the linkage between midbrain activation and memory formation have been recently reported. In a long-term memory study for object pictures Wittmann et al. (2005) reported greater dopaminergic midbrain activation associated with items that predicted monetary rewards and that these items were recalled later at a higher degree and were associated with greater hippocampal activation. A recent study by Adcock et al. (2006) demonstrated that long-term memory for scenes was influenced by monetary rewards and that the ventral tegmental area showed greater connectivity with the hippocampus suggesting a midbrain reward-related role in facilitating learning. These studies indicate that rewards can affect brain areas involved in facilitating the storage of information when reward-based information processing occurs.

In this study, we sought to investigate the effects of reward motivation on neural systems supporting higher levels of cognitive control. Recent neuronal recording studies from non-human primates have reported that neurons within the lateral PFC show firing to both reward motivation and abstract task demands. Hikosaka and Watanabe (2000) characterized the lateral PFC as being a site for the integration of cognition and reward and Watanabe et al., 2002a, Watanabe et al., 2002b have reported evidence of cognitive control and reward in the firing characteristics of lateral PFC neurons. Leon and Shadlen (1999) reported evidence of reward facilitation of lateral PFC neurons involved in a WM task. In this experiment task-related neurons showed a further amplification of firing rate when a desirable food trial-outcome was cued relative to when a less-desirable trial-outcome was cued. Similar results in non-human primate lateral PFC come from Kobayashi et al. (2002), who reported increases in neuronal firing rates to reward trials, but these investigators also reported populations of neurons tuned to the combination of target maintenance and non-reward. Additionally, Tsujimoto and Sawaguchi (2004) reported cells within the lateral PFC that appeared to represent response outcome in a WM task. Overall these results indicate that task-relevant PFC neurons can be enhanced or suppressed by the influence of the reward contingency of a WM trial.

Functional MRI evidence of reward motivation influencing performance has been reported recently in WM, attention, and motor tasks. Taylor et al. (2004) found an interaction within the PFC between WM for shapes and reward motivation. Greater activation has also been reported within frontopolar cortex (BA 10), when reward motivation interacts with performance on a WM n-back task (Pochon et al., 2002) and in verbal maintenance regions when reward is offered for performance on task requiring WM for words (Gilbert and Fiez, 2004). In a spatial attention task, parietal regions associated with the allocation of spatial attention in visual cueing task were enhanced by the presence of reward incentives for speeded performance (Small et al., 2005). Visual cortex also showed enhancement of signal when reward was present in that spatial-attention task. Lastly, motor response preparation was linked with reward by Ramnani and Miall (2003), who reported greater activation within the left parahippocampal gyrus when reward was present in their motor task. In summary, these prior studies demonstrate that cortical areas associated with cognitive performance can show enhanced BOLD responses when performance-based rewards can be obtained.

In the current study, we investigate the influence of reward on cognitive control in WM by examining modulatory responses in visual association cortex. Monkey studies of inferior temporal (IT) cortex during delayed-match-to-sample (DMS) tasks have demonstrated support for the modulatory influences of task goals on activity in this region during behavior. For example, Miller and Desimone (1994) trained monkeys to perform a DMS task in which they had to maintain and make match judgments about picture stimuli. In addition to finding IT neurons associated with the active maintenance of pictorial information, they also noted that a subset of the population of neurons showed enhancement of spiking activity at the time of match judgment when a previously shown image was repeated. Also, other neurons suppressed their firing rate when the test picture was a match. These neuronal tuning characteristics indicate that IT neurons are sensitive to task demands and may participate actively in the judgment process, possibly via connections with the lateral PFC (Fuster et al., 1985, Miller et al., 1993, Knight et al., 1999).

Mounting evidence from human studies also indicates that the lateral PFC exerts control over visual association areas (Barcello et al., 2000, Miyashita and Hayashi, 2000, Ranganath et al., 2004, Rose et al., 2005). For example, Gazzaley et al., 2005a, Gazzaley et al., 2005b used a delayed recognition task in which pictures of faces and scenes were presented serially at encoding and subjects were instructed to attend to one category and ignore the other. Both fMRI and ERP results supported a modulation of stimulus-relevant activity within IT visual association cortex. For scene stimuli, regions within the parahippocampal gyrus showed significant activation for the perception of scene stimuli. Activation within these regions was enhanced when scenes were to be attended relative to a perceptual baseline condition and suppressed relative to baseline when scenes were to be ignored. A similar modulation was reported for face-selective visual association cortex. The current experiment will build upon these findings by testing the effect of increased reward motivation on top-down modulation using this WM paradigm.

In this study, we present results from an fMRI experiment that pairs reward and WM using a variant of the task described by Gazzaley et al., 2005a, Gazzaley et al., 2005b in which pictures of faces and scenes are presented at encoding with the requirement that one category be attended to for later memory after a delay, while the other category is ignored. We investigated whether the top-down modulatory patterns of enhancement and suppression described previously in scene-selective parahippocampal gyrus and in a face-selective area of fusiform cortex are modified by reward motivation. Furthermore, we tested whether task-relevant regions of the PFC would also show this modulatory pattern with reward and WM. Additionally, we attempted to determine the extent to which either of these modulatory patterns influences performance of the task relative to non-rewarded task conditions.

Section snippets

Behavioral data

Overall accuracy on all trials was relatively high across all reward levels. Accuracy was uniformly high in the passive view trials (99.47% for low reward and 100.00% for high and non-reward), thus all analyses of mean accuracy included comparisons of the remember faces/ignore scenes and remember scenes/ignore faces only. We had predicted that rewards would potentially influence accuracy rates leading to higher accuracy when a reward was possible. Means for accuracy were analyzed with a 3

Discussion

We present evidence that reward motivation can influence a task that requires top-down attentional signaling to perform at optimal levels. Specifically, we found that incentives influence top-down attentional signals, that is, when rewarding goals are available, top-down modulation can be driven to higher levels, due to both increased ability to enhance task-relevant inputs and to suppress competing irrelevant perceptual inputs. This finding adds to a growing body of research emphasizing the

Subjects

Sixteen volunteer subjects (6 females) from the University of California participated after providing informed consent. Age ranged from 19 to 37 (M = 22.22, S.D. = 3.86). All subjects had normal or corrected vision, were free of neurological disorders, and were not taking any medications having a psychoactive, cardiovascular, or homeostatic effect. Fifteen subjects were right-handed and one was predominantly, but not exclusively left-handed as reported in our consent protocol.

Cognitive task and procedure

A WM delay recognition

Acknowledgments

We thank Rahul Modi and Ben Bendig for assistance with the task. We also thank Roshan Cools, Charlotte Boettiger, Christian Fiebach, Jeff Cooney, and members of the D'Esposito laboratory and Neural Circuits and Brain Imaging Program at the Gallo Clinic and Research Center for their helpful comments and suggestions.

References (48)

  • W. Schultz

    Getting formal with dopamine and reward

    Neuron

    (2002)
  • S.F. Taylor et al.

    A functional neuroimaging study of motivation and executive function

    NeuroImage

    (2004)
  • E.M. Tricomi et al.

    Modulation of caudate activity by action contingency

    Neuron

    (2004)
  • B.C. Wittmann et al.

    Reward-related FMRI activation of dopaminergic midbrain is associated with enhanced hippocampus-dependent long-term memory formation

    Neuron

    (2005)
  • E. Zarahn et al.

    A trial-based experimental design for fMRI

    NeuroImage

    (1997)
  • C.F. Zink et al.

    Human striatal responses to monetary reward depend on saliency

    Neuron

    (2004)
  • R.A. Adcock et al.

    Reward-motivated learning: mesolimbic activation precedes memory formation

    Neuron

    (2006)
  • F. Barcello et al.

    Prefrontal modulation of visual processing in humans

    Nat. Neurosci.

    (2000)
  • L.L. Chao et al.

    Contribution of human prefrontal cortex to delay performance

    J. Cogn. Neurosci.

    (1998)
  • A.R. Damasio

    The somatic marker hypothesis and the possible functions of the prefrontal cortex

    Philos Trans. R. Soc. Lond., B Biol. Sci.

    (1996)
  • M.R. Delgado et al.

    Dorsal striatum responses to reward and punishment: effects of valence and magnitude manipulations

    Cogn. Affect. Behav. Neurosci.

    (2003)
  • R. Elliott et al.

    Dissociable neural responses in human reward systems

    J. Neurosci.

    (2000)
  • J.M. Fuster

    Cortex and Mind

    (2003)
  • A. Gazzaley et al.

    Top-down enhancement and suppression of the magnitude and speed of neural activity

    J. Cogn. Neurosci.

    (2005)
  • Cited by (140)

    • Cerebellar and cortico-striatal-midbrain contributions to reward-cognition processes and apathy within the psychosis continuum

      2022, Schizophrenia Research
      Citation Excerpt :

      Moreover, loss of cognitive load effect after correction for premorbid IQ scores accords with evidence showing a positive association between high scores and better n-back task performance (Jaeggi et al., 2010). These findings are consistent with the previously established impact of cognitive load and a facilitatory role for reward in performance (Krawczyk et al., 2007, Braver et al., 2014, van den Berg et al., 2014, Etzel et al., 2016) and show that reward and cognitive load manipulations were overall effective. At the neural level, we found that the posterior midline cerebellum and the thalamus were implicated in processing both reward and cognitive load in the psychosis continuum.

    • Reward and loss incentives improve spatial working memory by shaping trial-by-trial posterior frontoparietal signals

      2022, NeuroImage
      Citation Excerpt :

      This heterogeneous effect of motivated working memory on prefrontal BOLD signals has previously been seen in other work, with relatively decreased BOLD signals in ventrolateral and ventromedial PFC, and BA10 previously reported (Pochon et al., 2002; Gilbert and Fiez, 2004; Longe et al., 2009). Our results most closely match studies of motivated object or visual WM, both of which typically require visual association cortex during baseline WM, and also when incentive information is integrated (Taylor et al., 2004; Krawczyk et al., 2007; Krawczyk and D'Esposito, 2013). In the case of the study by Taylor and colleagues, which used a delay response visual WM task, elevated BOLD activation was seen in the IPS and superior frontal sulcus, similar to our study (Taylor et al., 2004).

    View all citing articles on Scopus
    View full text