SEARCH

SEARCH BY CITATION

Keywords:

  • motivation;
  • motor-evoked potential;
  • movement;
  • transcranial magnetic stimulation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Much human behavior is driven by urges. Yet research into urges is hampered by a paucity of tools to objectively index their strength, timing and control. Here we used transcranial magnetic stimulation (TMS) and concurrent electromyography to examine whether urges for food and money are detectable via motor system excitability. In Experiment 1, we used a naturalistic food paradigm to show that food items that were most strongly wanted elicited the largest motor excitability, even before participants knew which response to make to get them. In Experiment 2a, we replicated the results using money – motor excitability was greater for larger monetary amounts. In Experiment 2b we show that monetary amount does not modulate motor excitability when participants simply observe, without having to take action. As the chief effect occurred prior to the subject knowing which motor response to make, it is not merely related to response preparation, and as the effect was present only when action was required, it is not merely related to increased arousal. Instead, the increased motor excitability likely indexes the degree of motivation a subject has to perform an action. Thus, we have used TMS to demonstrate that urges for food and money ‘spill over’ into the motor system. This is likely mediated by interactions between the limbic system (including the orbital frontal cortex) and the motor system, probably at the level of the basal ganglia. Implications are discussed for theories of embodied cognition and for methodological progress in studying urge control.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

While some kinds of impulses are mainly action-oriented (e.g. the impulse to step into the street when the light changes), other kinds are more motivational (e.g. the impulse for a cigarette). We refer to this latter form of impulse as an ‘urge’. It relates to how much someone wants something, driven by its perceived value. Urges constitute an important part of human behavior, both in healthy everyday life and in psychiatric disorders. Yet there is a paucity of methods to objectively index urges in terms of strength, timing (dynamics) and control. While it is possible to measure the strength of the urge in terms of response time, or number of items chosen/consumed, or subjective self-report (Raylu & Oei, 2004; Seibt et al., 2007; Wulfert et al., 2009), these behavioral measures do not provide information about the dynamic unfolding of the urge in real-time, nor are they suitable for measuring urge control. If an urge is stopped then there is nothing to observe behaviorally.

We aimed to develop a technique to measure urges by assuming they would ‘spill over’ into the motor system. This assumption has a precedent. For example, it has been shown that action is more vigorous for stimuli with higher motivational value, and that this has its counterpart in increased blood oxygen level-dependent (BOLD) activation in the nucleus accumbens area (Talmi et al., 2008). A different study used functional magnetic resonance imaging (fMRI) and skin conductance to show that value modulates behavioral activation and BOLD signal in the pallidum even with subliminal stimuli (Pessiglione et al., 2007). Yet a limitation of these studies is that the subject knows exactly which response to make, so the increased activation may also reflect motor execution rather than a purer measure of motivation to respond. Nor do these measures provide the sub-second resolution needed to separate the effects of motivation from those of execution.

A different approach used transcranial magnetic stimulation (TMS) of the primary motor cortex to show that motor excitability (recorded from the hand) was modulated by an upcoming potential reward (Kapogiannis et al., 2008). However, that study required passive viewing without any action and, moreover, varied both reward value and the probability of getting reward, thus making it unclear whether the increased motor excitability relates to urge per se rather than any of arousal, expectancy or uncertainty.

We developed a novel approach to index urges in the motor system using TMS and concurrent electromyography. In Experiment 1 we used a realistic and previously validated food paradigm with hungry human participants (Hare et al., 2009). In Experiment 2 we used a similar paradigm with monetary rewards. We hypothesized that stimuli associated with stronger urges (for food or money) would lead to higher motor excitability. We aimed to show that this would be manifest even before the subject knew which motor response to make. We also aimed to clarify the within-trial timing of the effect and also to address whether the effect depends on making an action.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Participants

For Experiment 1, there were 17 young adult participants (nine male; all right-handed; mean age = 21.9 years, range 18–26 years). One participant did not complete the study because of technical problems with the acquisition system – this person’s data are not included. Participants were instructed to not eat for 4 h prior to the experiment. For Experiment 2, 15 young adults participated in versions 2a and 2b in one overall session in counterbalanced order (eight male; one left-handed; mean age = 20.4 years, range 18–26 years). All participants provided written consent in accordance with the Internal Review Board guidelines of the University of California at San Diego. Participants also completed a TMS safety-screening questionnaire and were found to be free of contraindications.

Experiment design

Experiment 1: the food paradigm with action

The paradigm was based on Hare et al. (2009). Sixty food items were placed in a box in the experiment room. The items comprised a mix of appetitive items (e.g. candy bars) and (generally) aversive items (e.g. clam juice). Participants also viewed digital images of all food items on the computer to familiarize themselves with the items before rating them. Each food item was then presented on the screen, one by one, and participants rated the item on a five-point scale (‘Sure-No’, ‘Probably-No’, ‘Neutral’, ‘Probably-Yes’, ‘Sure-Yes’), indicating if they would like to eat the item at the end of the experiment. These five rating levels were interpreted as five urge levels in our analysis: strongly unwanted, weakly unwanted, neutral, weakly wanted and strongly wanted.

Before beginning the main experiment, participants performed a short practice session of eight trials. Participants subsequently performed a total of four blocks of 70 trials, with each block containing 60 ‘food trials’ and 10 ‘blank trials’. Thus, each food stimulus was repeated four times. The order of stimuli was randomized within each block. Each trial began with a cue (a picture of food, or an empty rectangle for blank trials) for 2 s, followed by a blank screen for 1 s (Fig. 1A). A choice screen followed, showing [Yes No] or [No Yes], selected randomly, for up to 1 s, during which time the participant made a response with the left or right index finger, depending on whether she wanted to eat the item. Thus, participants had to wait until the appearance of the choice screen to know which hand was needed to make the appropriate response. On each trial, a TMS pulse was delivered at only one of the two time-points: ‘early’ (1.5 s before the choice screen) or ‘late’ (0.5 s before the choice screen), with 50% of the trials getting each type of pulse. For blank trials, participants were instructed that it was immaterial whether they select YES or NO, but they must make one of the two responses. There was a 2-s inter-trial interval (ITI). Participants were informed that, at the end of the experiment, one of the trials would be randomly selected and honored (i.e. participants got to eat the item if they selected Yes, but did not get to eat the item if they selected No), and therefore the best strategy was to make an accurate selection on every trial, see Hare et al. (2009).

image

Figure 1.  (A) Task design for Experiment 1. The cue was a picture of food, or an empty rectangle for blank trials. It was followed after 2 s by a blank screen. A choice screen followed after 1 s, showing [Yes No] or [No Yes], for up to 1 s, during which time the participant made a response with the left or the right index finger, depending on whether she wanted to eat the item. For blank trials, participants were instructed that it was immaterial whether they select YES or NO, but they should make one of the two responses. The inter-trial interval was 2 s. A transcranial magnetic stimulus (TMS) pulse was delivered either at an ‘early’ time-point (1.5 s before the choice screen) or at a ‘late’ one (0.5 s before the choice screen). (B) TMS setup. MEPs were recorded from the right FDI muscle with TMS being delivered over the left M1 using a figure-of-eight coil. (C) TMS results. A strong urge to consume the food results in higher MEPs at the late but not at the early stimulation period. Error bars show SEMs.

Download figure to PowerPoint

Experiment 2a: the money paradigm with action

Participants performed one block of 60 trials, comprising 40 ‘money trials’ and 20 ‘blank trials’, presented in randomized order. Money trials included 20 repetitions of a $5 bill and 20 repetitions of a 10 cents coin. Each trial began with a cue (a picture of a $5 bill or a 10 cents coin within a white rectangle; or an empty rectangle for blank trials) for 2 s, followed by a blank screen for 1 s (Fig. 2A). TMS was delivered at only one time-point – this was 500 ms before the choice screen (like the ‘late’ period of Experiment 1). This was motivated by the finding in Experiment 1 (see below) that this time-point was the optimal one for eliciting an effect. After this, a choice screen appeared (as for Experiment 1) and the participant selected the response. Some trials included a yellow border around the white rectangle during money stimulus presentation; on these trials, the participant was required to say ‘yellow’ (see Experiment 2b below). Participants were informed that, at the end of the experiment, one of the money trials would be randomly selected and honored (i.e. participants get the money if they selected Yes). Participants were instructed to select Yes on both types of money trials (optimal choice for participants), as well as on the blank trials. Having the same response on all three types of trials ensured that any resulting differences in motor-evoked potentials (MEPs) across these trials were dependent only on the monetary value of the trials, and not independently driven by differences in the required responses.

image

Figure 2.  (A) Task design for Experiments 2a and 2b. The cue was a picture of a $5 bill or a 10 cents coin within a white rectangle; or the cue was an empty rectangle for blank trials. This was followed after 2 s by a blank screen for 1 s. After this, a choice screen appeared (as for Experiment 1) and the participant responded. A minority of trials included a yellow border around the white rectangle during money stimulus presentation; on these trials, the participant was required to say ‘yellow’. TMS was delivered 500 ms before the choice screen (same as ‘late’ time-point for Experiment 1). (B) TMS results for Experiment 2a. A strong urge results in higher motor-evoked potential (MEPs; when responses are required). (C) TMS results for Experiment 2b. A strong urge does not result in higher MEPs (when responses are not required). Error bars show SEMs. * denotes statistical significance.

Download figure to PowerPoint

Experiment 2b: the money paradigm without action

The task structure was similar to Experiment 2a (Fig. 2A), the only differences being that the choice screen was not presented and the participants did not have to move their fingers to press keys. To minimize the possibility of participants not paying attention to the screen (as no hand responses were required in this experiment), each trial included, with a 10% probability, a yellow border on the white rectangle containing the money cue. On these trials, the participants had to say the word ‘yellow’ as soon as they saw the border; they were instructed that failure to do so more than once would result in the cancellation of any monetary rewards they might otherwise receive from the experiment. To keep Experiments 2a and 2b similar, this additional feature and requirement was also included in Experiment 2a. Note that Experiment 2b did not have any manual response requirement. The only motor requirement was to report the occurrence of the yellow border on the 10% of trials in which this occurred.

Apparatus

Participants were seated 50 cm in front of an iMac (19-inch monitor). The experiments were run using Matlab (MathWorks, Natick, MA, USA) and the PsychToolBox3 (http://www.psychtoolbox.org). In the pre-TMS phase of Experiment 1, participants used five keys (G, H, J, K, L) on the keyboard to indicate their urge level for food items, from the lowest to the highest. In the TMS phase of all experiments, participants sat with their forearms resting on the chair armrest and the table surface in front of two keypads (19-key numeric keypad; Adesso, Walnut, CA, USA). Participants placed the index finger against a key on the vertically placed keypad such that they could respond with a key press by moving the finger inward in a lateral abduction. This lateral movement of the finger is necessary to isolate the index finger muscle for electromyographic (EMG) recording (see below).

EMG recordings

Surface EMG recordings were made via 10-mm-diameter Ag–AgCl hydrogel electrodes (Medical Supplies, Newbury Park, CA, USA) placed over the right first dorsal interosseous muscle (FDI – index finger). Ground electrodes were placed over the styloid process of the right radius. The EMG signal was amplified using a Grass QP511 Quad AC Amplifier System Grass amplifier (Grass Technologies, West Warwick, RI, USA), with a band-pass filter between 30 Hz and 1 kHz and a notch filter at 60 Hz. Data were sampled at 2 kHz using a CED Micro 1401 mk II acquisition system, and displayed and recorded to disk using CED Signal v4 (Cambridge Electronic Design, Cambridge, UK).

TMS

We used a MagStim 200-2 system (MagStim, Whitland, UK) with a figure-of-eight coil (7-cm diameter) to deliver a single test stimulus during task performance (Fig. 1B). The coil was positioned to produce the largest, reliable MEPs in the right FDI. Resting motor threshold was determined by finding the lowest stimulus intensity that produced MEPs of at least 0.05 mV amplitude on at least five of 10 trials (Rossini et al., 1994). Test stimulus intensity was set to about 110% of the resting motor threshold, as this level was found to produce an MEP that was approximately half of the participant’s maximum MEP amplitude. This ensured that the test stimulus intensity was on the ascending limb of the individual’s stimulus–response curve, so that both increases and decreases in corticomotor excitability could be detected (Devanne et al., 1997).

Analysis

Each trial provided an MEP measurement for the FDI muscle. In Experiment 1, MEPs were categorized as ‘early’ or ‘late’, depending on the timing of the stimulation. MEPs from food trials for the two time-points were normalized by dividing by the average MEP from blank trials for that time-point. MEPs for early and late categories were further grouped into five urge levels, depending on the rating given by the participant in the pre-TMS phase of the study. In Experiments 2a and 2b, MEPs from money trials were normalized by dividing by the average MEP from blank trials. MEPs were grouped into two urge levels, strong ($5 trials) and weak ($0.1 trials). In all experiments, MEPs in each urge level were 10% winsorized, i.e. the smallest and the largest 10% of the MEPs were set to the MEPs at the 10% and the 90% percentile boundary, respectively. Tables 1 and 2 show the raw MEPs and reaction times (RTs) for the different conditions of the three experiments. The primary analysis was performed on baseline-normalized rather than raw MEPs (as the variance was generally smaller for the former). Analysis used PASW Statistics 17.0.2 (SPSS, Chicago, IL, USA). In a verification procedure, we computed root mean square pre-TMS EMG activity for each condition in order to establish if the muscle of interest was at rest at the time of stimulation.

Table 1.   Behavioral and raw TMS results from Experiment 1
 Early MEP (mV)Late MEP (mV)RT (ms)
  1. Mean raw motor-evoked potentials (MEPs) and reaction times (RT) are shown for the five categories (based on consumption urge) of food trials and for the blank trials (baseline). The first two columns represent MEPs for the early and the late stimulation periods. Values inside parentheses indicate standard deviation.

Baseline0.97 (0.35)0.93 (0.34)502 (97)
Neutral0.91 (0.34)0.81 (0.28)551 (70)
Weakly wanted0.86 (0.32)0.85 (0.34)531 (63)
Strongly wanted0.89 (0.29)0.98 (0.43)528 (78)
Weakly unwanted0.85 (0.30)0.88 (0.33)533 (66)
Strongly unwanted0.84 (0.36)0.87 (0.35)539 (79)
Table 2.   Behavioral and raw TMS results from Experiments 2a and 2b
 Experiment 2aExperiment 2b
MEP (mV)RT (ms)MEP (mV)
  1. Mean raw motor-evoked potentials (MEPs) and reaction times (RT) are shown for $0.1 (weakly wanted), $5 (strongly wanted) and the blank (baseline) trials. Values inside parentheses indicate standard deviation.

Baseline0.78 (0.23)497 (94)0.53 (0.20)
Weakly wanted0.69 (0.21)479 (86)0.54 (0.16)
Strongly wanted0.77 (0.27)470 (65)0.55 (0.19)

The key hypothesis in Experiment 1 was that MEPs would increase with urge. Accordingly, we tested whether there was a linear increase from neutral to weakly wanted to strongly wanted items at early and late time-points separately. The same analysis was also done for RT. In addition, we used aversive food stimuli to both increase the range of subjective urge measurements and to examine the relationship between MEP and ‘negative’ urges (i.e. motor system responses for items the participant did not want to consume). We did not have any prediction about how MEPs would relate to the strength of the negative urges. The key hypothesis for Experiment 2a was that MEPs would be greater for the $5 stimulus than the 10 cent stimulus. For Experiment 2b, we were interested to see if the absence of action would produce the same or different results from Experiment 2a.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

For Experiment 1, a linear contrast across wanting levels showed that normalized MEPs increased significantly with increasing urge for the late period (F1,15 = 6.536, P = 0.022), but not for the early period (F1,15 = 0.191, n.s.; Fig. 1C). Post hoc, Bonferroni-corrected tests for the late period showed a significant difference in normalized MEPs between the strongly wanted and the neutral conditions (t15 = 2.557, P < 0.0167), and for the strongly wanted and the weakly wanted conditions (t15 = 2.371, P < 0.0167), but not for the weakly wanted compared with the neutral condition (t15 < 1). These effects remained unaltered when raw (rather than baseline-normalized) MEPs were used in the analysis.

A linear contrast across wanting levels also showed that RT got faster with increasing consumption urge (F1,15 = 8.072, P = 0.012). Post hoc, Bonferroni-corrected tests revealed a significant decrease in RT for the strongly wanted compared with the neutral condition (t15 = 2.841, P < 0.0167), and for the weakly wanted compared with the neutral condition (t15 = 2.619, P < 0.0167), but not for the strongly wanted compared with the weakly wanted condition (t15 < 1). A verification analysis of root mean square pre-TMS EMG activity showed that the muscle was equally at ‘rest’ for the comparison of strongly wanted and neutral conditions (t15 < 1, n.s.).

To analyse ‘negative urges’, for which we had no predictions, we performed a repeated-measures anova for the neutral, the weakly unwanted and the strongly unwanted conditions for the early and late periods separately. This did not reveal any significant effect of ‘negative urges’ on normalized MEPs, for the early (F2,30 = 2.35, n.s.) or the late (F2,30 < 1, n.s.) stimulation periods.

In Experiment 2a, the key hypothesis was that MEPs would be higher for the strong urge condition ($5 trial) compared with the weak urge condition ($0.1), parallel to the effect already seen in Experiment 1. The results showed the expected effect of higher MEPs in the $5 condition (t14 = 2.085, P = 0.028; Fig. 2B). Correspondingly, the average RT was smaller in the strong urge condition compared with the weak urge condition by 9 ms, although the difference was not statistically significant (t14 < 1). A verification analysis of root mean square pre-TMS EMG activity showed that the muscle was equally at ‘rest’ for these conditions (t14 = 1.3, n.s.).

In contrast, the MEPs in Experiment 2b for the strong urge condition were not found to be larger than the MEPs for the weak urge condition (t14 = −0.178, n.s.; Fig. 2C). A verification analysis of root mean square pre-TMS EMG activity showed that the muscle was equally at ‘rest’ for these conditions (t14 < 1, n.s.). In both Experiments 2a and 2b, on the 10% of trials in which the yellow border was presented, participants satisfactorily reported this occurrence (< 2 errors for each subject). This showed that the subjects were paying attention to the stimuli in Experiment 2b even though no manual motor response was required.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We recorded TMS-induced MEPs from the right index finger to measure the level of urges for food and money. In Experiment 1, using the food paradigm, we found that MEPs increased with increasing urge for food, specifically at the late but not at the early time-point. Importantly, these measurements were made before the participant even knew which motor response to make. The effect was replicated for money (Experiment 2a), proving reliability and generalizability. Next, by removing the response requirement (Experiment 2b), we show that a critical element of the ‘urge effect’ measured here is the need for subjects to take action.

Our results agree with the findings of Pessiglione et al. (2007) who showed that even subliminal high-value stimuli lead to behavioral and BOLD activation. However, because that study relied on (slow) fMRI measures, it could not dissociate the preparation of movement from the actual movement. Here, the high temporal resolution of TMS points to urge-related motor excitability before movement. Moreover, in the study by Pessiglione et al. (2007), when the stimulus occurred participants knew which response to prepare. In our experiment, we changed the mapping of Yes and No choices to left and right hand randomly on each trial, and thus recorded MEPs before the participants even knew which response they would need to make. By doing this, we rule out the possibility that the observed increases of MEPs merely relates to the preparation of a particular motor response. Instead, it must also reflect motivational processing that is upstream from the corticospinal system. Such upstream processing likely involves brain systems such as the orbital frontal cortex (activated in this same paradigm parametrically with wanting of food items; Hare et al., 2009) and associated limbic circuitry, including the ventral striatum/nucleus-accumbens and ventral pallidum (Berridge et al., 2009). Our observation that motivation ‘spilled over’ into the motor system could have its neural counterpart in communication between ‘limbic’ and ‘motor’ loops at the level of the basal ganglia (Joel et al., 2002; McHaffie et al., 2005). fMRI could be used to test the prediction that motivational ‘spill over’ will correspond to increased activation of motor territories of the basal ganglia, even before action ensues.

Our ability to measure the urge before action ensues also has strong advantages over prior studies that have tried to measure the strength of an urge in terms of response time, or number of items chosen/consumed, or subjective self-report (Raylu & Oei, 2004; Seibt et al., 2007; Wulfert et al., 2009). These behavior-based studies provide readout of the motor system only after the action is made, making them unsuitable for studies of urge control (in which there is no behavior to observe), and for studies of urge dynamics (the timing of urge formation and the factors that affect it).

The urge-related signal we detected in the motor system, before action, may be interpreted in the framework of an expanding literature called ‘embodied cognition’. Many results across language, emotion and decision-making are being interpreted in terms of the ‘spill over’ of a cognitive process (e.g. an urge, decision or thought) into the motor system (e.g. Barsalou, 1999; Gold & Shadlen, 2000; Pulvermüller, 2005; Semin & Smith, 2008). Of specific relevance, some recent studies have used 3D movement tracking to show how perceptual, cognitive and linguistic decisions may spill over into an executed motor movement even before the decision has been fully completed (Spivey et al., 2005; Song & Nakayama, 2009). Our results are complementary to these findings. However, they have the strength of providing a sensitive and readily acquired neurophysiological measure even before the subject knows which action to take. Thus, the TMS method may be particularly well suited to capturing ‘spill over’ of motivation onto the motor system, even before the motor system knows precisely what to do.

This study highlights the importance of stimulation timing. In the food paradigm, the effect of urge on MEPs was visible 500 ms before the choice, but not 1500 ms before. Clearly a better understanding of the temporal dynamics of this influence will require analysis of MEPs at many more than two time intervals; however, the current study provides a starting point for an informed selection of appropriate intervals.

By comparing Experiments 2a (action required) and 2b (no action required) we show that a critical element of the ‘urge effect’ is the necessity to take action to get the reward. This is a useful finding because it shows that the ‘urge effect’ measured with TMS is not merely related to increased general brain arousal (e.g. to seeing $5 as opposed to 10 cents), but also reflects something about action. We started out defining an ‘urge’ as ‘how much someone wants something’. Based on the current paradigms at least, the concept of urge could be refined to ‘how much someone wants something when action is required to get it’.

In respect of the need for action, our findings are at odds with those of (Kapogiannis et al., 2008), who identified an effect of reward on the motor cortex using paired-pulse TMS in a paradigm in which the participants did not make any response. However, with the task used in their study, they could not identify whether the effect was determined by the size of the reward or the probability of receiving it (in that study the effect related to a strong reduction in uncertainty when observing whether a reward materialized over a specific time interval). We suspect that the changing reward probabilities drove the observed effect and, therefore, having an action was not critical in their paradigm. Here, in our money paradigm, the task was set up to measure the strength of the urge, manipulated solely by the size of the monetary reward ($5 or $0.1), while keeping the probability of seeing $5 or $0.1 on any trial exactly the same.

In these experiments, the MEPs likely reflect multiple contributing factors, including not only the urge (determined by the value of the stimulus), but also action preparation. Hence, we caution the reader that comparing MEPs (raw or normalized) across different experiments might be misleading. It is only the relative difference between MEPs observed for different levels of urge (within an experiment, when all other factors are controlled) that can be reliably interpreted. Even a comparison of MEPs between baseline and non-baseline trials within the same experiment is difficult to interpret because the probability of seeing a baseline trial was lower than the probability of seeing a non-baseline trial in all three experiments and, therefore, differences in the probabilities could confound such a comparison. We used baseline trials only for normalizing the MEPs within each participant (to reduce variance between participants).

Thus, we have shown that the strength of an urge can be indexed via ‘spill over’ into motor system excitability, at one time-point and not another, and only when a response is needed for satisfying the urge. Moreover, unlike prior studies, we have separated the preparation to make a response from response execution itself. Further, by recording motor excitability before the participant knew which response to prepare, we also show that the effect on motor excitability is not purely one of motor preparation but must also reflect a motivational component. Further, by manipulating the response-requirement in the money task, we show that the effect is also not purely related to general brain arousal but must also include an action-relevant component.

The methods developed here could have useful applications. It may be possible to harness the high temporal resolution of TMS to address the dynamics of how urges rise and fall when cognitive control is applied. For example, by delivering TMS pulses at specific time-points on NoGo trials in a Go/NoGo paradigm (Yamanaka et al., 2002) or on stop trials in stop signal paradigms (Coxon et al., 2006; van den Wildenburg et al., 2008) it is possible to visualize how response activation is followed by response suppression. A similar methodology could be used to examine how ‘urge’ activation is suppressed when cognitive control mechanisms are applied. Such studies could show whether failures in urge control, such as those occurring in many psychiatric disorders, are due to excessive motivation or poor control, or both.

The current study grounds motivation in the motor system. This leads to neuroscience predictions that could be verified with functional imaging and other methods. For example, it will be interesting to examine if motivational ‘spill over’ corresponds to increased activation of motor territories of the basal ganglia. It will also be interesting to examine whether cognitive control that is targeted at the motor system, for example via fronto-striatal or fronto-subthalamic inputs, could diminish motivation.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Antonio Rangel of Caltech for sharing the food stimuli with us and for instructing us in setting up the behavioral paradigm. We thank Piotr Winkielman for helpful comments on the manuscript. We gratefully acknowledge support from the Alfred P. Sloan Foundation and NIH NIDA Grant DA026452 to A.R.A. (PI).

Abbreviations
BOLD

blood oxygen level-dependent

EMG

electromyogram

FDI

first dorsal interosseous muscle

fMRI

functional magnetic resonance imaging

MEP

motor-evoked potential

RT

reaction time

TMS

transcranial magnetic stimulation

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References