Wednesday, May 06, 2009

The Constant State of Desire

Karen Finley in The Constant State of Desire
Photo by Donna Ann McAdams

If you are better at exerting self-control by choosing less pleasurable but more healthy options, do you live in a constant state of (suppressed) desire? If you feel like you're always sacrificing and denying yourself what your brain's "valuation system" indicates is worth more, does the craving eventually go away? Or are you more likely to binge in a moment of weakness? Isn't this why most diets fail? Wouldn't it be better to have that candy bar every once in a while? (or else, learn to love cauliflower more than Cadbury)?

However, this constant state of desire is not what the mad lib Science paper (Hare et al., 2009) is about. It's not even about dieting at all, despite what the press release says.
Caltech Researchers Pinpoint the Mechanisms of Self-Control in the Brain

Study of dieters shows how two brain areas interact in people with the willpower to say no to unhealthy foods

PASADENA, Calif.--When you're on a diet, deciding to skip your favorite calorie-laden foods and eat something healthier takes a whole lot of self-control--an ability that seems to come easier to some of us than others. Now, scientists from the California Institute of Technology (Caltech) have uncovered differences in the brains of people who are able to exercise self-control versus those who find it almost impossible.

The key? While everyone uses the same single area of the brain to make these sorts of value-laden decisions, a second brain region modulates the activity of the first region in people with good self-control, allowing them to weigh more abstract factors--healthiness, for example--in addition to basic desires such as taste to make a better overall choice.
To start, let's look at the subject selection criteria (Hare et al., Supporting Online Material - PDF):
"We recruited two types of subjects: 1) individuals who self-reported being on a diet to lose or maintain weight and 2) individuals who self-reported no current monitoring of their diet. All subjects reported that they enjoyed eating sweets, chocolate, and other “junk food” even though they might be restricting them from their current diet.

. . .

Subjects were classified as self-controllers (SC) or non-self-controllers (NSC) based on their behavior during the experiment, and not on their self-reports about diet status during the recruiting process."
So really, the study isn't about dieting at all,1 because one could be a "self-controller" in the experiment and yet report no dietary restrictions in real life. And actually, the study may not even say much about food choices in the real world, because a participant could say she'd choose the broccoli over the brownie to gain the experimenter's approval, but then go home and eat ice cream for dinner.

Back to the subject selection:
"Fifty-two subjects participated in the experiment. However 15 subjects did not meet our a priori inclusion criteria based on their behavioral data [they fell into a gray zone]. 37 subjects were included in the analysis. Subjects were divided into two groups based on their behavioral data: successful self-controllers (SC) and non-self-controllers (NSC). The SC group included 19 subjects (14 female ... mean BMI = 24.8 ± 5.2) and the NSC group included 18 subjects (6 female ... mean BMI = 23.2 ± 5.1)."
There's a lot more pressure on women to diet and to give the appearance of restraint, and 70% of them were in the SC group. So who knows, it might be true that the female subjects felt more compelled to choose the healthy choice than they would have in real life.

Anyway, the experimental design is illustrated below:

Fig. 1A. The task proceeded in three parts: taste ratings, health ratings, and decisions.

The subjects viewed pictures of 50 different food items and rated them on a 5 point scale for health and taste in two separate blocks. After this, one food item rated as neutral on both dimensions (e.g., crackers, jello) was selected as the reference item, and subjects had to choose between it and another item on each trial.
"Subjects cared about their choices because they were required to eat the food that they chose in a randomly selected trial at the end of the experiment.2 Note that because subjects did not know which trial would count, their optimal strategy was to treat each decision as if it were the only one that counted. Although this is a binary decision task, subjects were asked to express the strength of their preferences using a five-point scale: Strong No (=choose reference) [to] Strong Yes (=choose shown item)."
The participants were sorted into SC and NSC groups according to their behavioral performance during the decision phase. The SC subjects chose food items on the basis of both health and taste, but the NSC group chose on the basis of taste alone. What were the investigators looking for in the fMRI data? They assumed at the outset that the brain's "value assigner" is located in the ventromedial prefrontal cortex3 (vmPFC), and the "self-controller" (in the dorsolateral prefrontal cortex, DLPFC) modulates value signals in vmPFC.

For fear of beating a dead horse, I won't get into the fMRI data analysis methods here (5 pages in the supplementary methods) and whether there were any "Voodoo Correlations" (or rather, "Puzzlingly High Correlations" - PDF). But here's a figure from Hare et al. with r=.847.

Fig. 2E. Robust linear regression showing a strong relationship between a measure of the effect that health ratings have on vmPFC activity and a measure of the effect that the health ratings have on decisions.

Below is another figure showing that Beta values in vmPFC increased with goal values, regardless of self-control. The legend was taken from Hare et al., Supporting Online Material (PDF).

Fig. 2B (Left). Estimated betas in the vmPFC for each of the regressors. This plot was constructed as follows. First, for each individual and type of trial we measured the associated beta value at the peak voxel for the goal value contrast (GLM 1) inside the vmPFC ROI shown in Fig. 2A (Right). The individual subject peaks were selected from within this ROI to allow for variability between subjects. Second, the mean and standard error of these betas were computed for each type of trial.

Now what about self-control in the DLPFC? A region in the left middle and inferior frontal gyri (IFG/BA 9) was more active during "self-control" trials in both groups, with the SC group showing this to a greater extent than the NSC group. And finally there was the requisite functional connectivity analysis (PDF) to explain the inverse relationship between greater activation in IFG/BA 9 and reduced activation in vmPFC. But since these two regions did not exhibit task-related functional connectivity, another area (IFG/BA 46, of course) had to be mediating the self-controlling effect of IFG/BA 9 on vmPFC. Got that? Hmm.

One of the things that struck me about this paper4 is the remarkable lack of scholarship, particularly when concluding in such a grandiose fashion:
"Lastly, an improved understanding of the neurobiology of self-control in decision-making will have applications to clinical practice in domains such as obesity and addiction, to economic and public policy analysis in problems such as suboptimal savings and health behaviors, and to legal thinking about which criteria should be used in determining if an individual is in full command of his decision-making faculties and thus accountable to the law."
Let's take the applications for obesity as our most relevant example here. Were there any citations of the literature on neurodegenerative disorders and appetite? Case studies of brain lesions and appetite? Studies of obese vs. lean individuals? Eating disordered vs. normal weight people? No, no, no, and no. Since they didn't cite any papers in these important [and more ecologically valid, if you really want to learn about self-control over food choices] areas of inquiry, here's what I learned from a PubMed search. A voxel-based morphometry study in patients with frontotemporal dementia demonstrated that binge eating was associated with greater degeneration in the right ventral insula, striatum, and orbitofrontal cortex (Wooley et al., 2007). Conversely, tumors in the right lateral frontal cortex have been associated with anorexia in several case reports (Houy et al., 2007; Trummer et al., 2002). In healthy subjects, transcranial direct current stimulation over DLPFC (specifically to increase neuronal firing over the right and to decrease it over the left DLPFC) was shown to reduce food cravings (Fregni et al., 2008). Neuroimaging studies have observed less activation in the left DLPFC of obese participants after a meal than was observed in lean participants (Le et al., 2006, 2007). There's even a right PFC hypothesis of obesity (Alonso-Alonso & Pascual-Leone, 2007).

You'd think all of this might have been relevant to an article on the neurobiology of self-control over food choices, but apparently not. Time for some chocolate cake. I just can't help myself...


1 The writer of the press release can be excused for getting this wrong because in the main article, Hare et al. lead the reader to believe that the subjects are all dieters:
To test these hypotheses, we recruited self-reported dieters and used functional magnetic resonance imaging (fMRI) to study the neural activity in vmPFC and DLPFC while the participants made real decisions about which foods to eat.
2 At the end of the experiment, were the participants asked how much they really cared about their choices, or is that just a proclamation?

3 The words dopamine, nucleus accumbens, and basal ganglia did not appear in the article (ignoring previous studies on the topic). The phrase "vmPFC-striatal network" appears once but is dissed right away.

4 ...besides everything else I've said thus far...

Additional Reading

Alonso-Alonso M, Pascual-Leone A. (2007). The right brain hypothesis for obesity. JAMA 297:1819-22.

Fregni F, Orsati F, Pedrosa W, Fecteau S, Tome FA, Nitsche MA, Mecca T, Macedo EC, Pascual-Leone A, Boggio PS. (2008). Transcranial direct current stimulation of the prefrontal cortex modulates the desire for specific foods. Appetite 51:34-41.

Houy E, Debono B, Dechelotte P, Thibaut F. (2007). Anorexia nervosa associated with right frontal brain lesion. Int J Eat Disord. 40:758-61.

Le DS, Pannacciulli N, Chen K, Del Parigi A, Salbe AD, Reiman EM, Krakoff J. (2006). Less activation of the left dorsolateral prefrontal cortex in response to a meal: a feature of obesity. Am J Clin Nutr. 84:725-31.

Le DS, Pannacciulli N, Chen K, Salbe AD, Del Parigi A, Hill JO, Wing RR, Reiman EM, Krakoff J. (2007). Less activation in the left dorsolateral prefrontal cortex in the reanalysis of the response to a meal in obese than in lean women and its association with successful weight loss. Am J Clin Nutr. 86:573-9.

Trummer M, Eustacchio S, Unger F, Tillich M, Flaschka G. (2002 ). Right hemispheric frontal lesions as a cause for anorexia nervosa report of three cases. Acta Neurochir (Wien). 144:797-801

Woolley JD, Gorno-Tempini ML, Seeley WW, Rankin K, Lee SS, Matthews BR, Miller BL. (2007). Binge eating is associated with right orbitofrontal-insular-striatal atrophy in frontotemporal dementia. Neurology 69:1424-33.


Hare, T., Camerer, C., & Rangel, A. (2009). Self-Control in Decision-Making Involves Modulation of the vmPFC Valuation System. Science, 324 (5927), 646-648. DOI: 10.1126/science.1168450

Introduction: The concept of self-control in decision-making has occupied philosophers and scientists throughout recorded history because the ability to exercise it is central to human success and well-being. Behavioral studies have examined the problem of self-control and provided valuable insights that suggest it is exhaustible in the short term (1–3), can be enhanced by cognitive strategies (4–7), and is correlated with measures of intelligence (8–10). However, little is known about the neurobiological underpinnings of self-control and how these neural mechanisms might differ between successful and unsuccessful self-controllers.

I told myself it would be different when I had children. We'd share our experiences and feelings together. We'd be so close. But I'm just like my father, a drunken slob. And the only feelings I share are no, no feelings at all. The only feelings I share are no feelings at all.

-Karen Finley, The Constant State of Desire (from Shock Treatment)

Subscribe to Post Comments [Atom]


At May 06, 2009 8:48 AM, Anonymous JohnP. said...

I'm not really clear on what's being criticized here. The sensational headlines or the paper?

I find the analysis strategy a little convoluted (to be fair I skimmed through the suplemental methods). But that's often the case with short science papers, there just isn't room for all the details. Also, the connectivity analyses struck me as interesting. A third region, often involved in regulation tasks, is mediating the relationship between BA9 and vmpfc. The authors walk us through how they found this region. They're not pulling a fast one or hiding anything. It's clear that they fished a bit, but as long as that's explained at the outset then I have no problem with it. In fact I encourage it, it's one of the ways we generate new hypotheses.

It's worth keeping in mind that, for better or worse, this is a science report. There are world limits and reference limits (30) Sure we can always nitpick and find something they forgot to cite, but they only had 30 refs to play with.

All that being said, the correlations do seem "surprisingly" high.

At May 06, 2009 8:27 PM, Blogger The Neurocritic said...

In this post, I'm more critical of the paper itself than of the press coverage. The "Self-Control Pinpointed" headline was overstated as usual and the self-control/dieting/willpower angle is certain to attract pop sci readers. But I thought the authors were a little misleading about their subject selection in the main part of the paper, because both dieters and non-dieters participated.

I didn't want to get into their analysis strategy here, because that would be a long post in itself. Instead, see Vul et al. 2009 and Kriegeskorte et al. 2009. In my view, it's a problem if readers don't know how the data were analyzed. For instance, it's not clear how the vmPFC ROI(s) were selected for the Fig. 2B/2E analyses (since there were 3 separate vmPFC ROIs). Science papers are short, but the Supporting Online Material for this paper was 20 pages long. Also for the conjunction analysis, the IFG/BA46 seed was not significant at p<.05 corrected.

As for the reference limits, I would have dropped the monkey physiology citations and some of the DLPFC cognitive control/working memory citations in favor of those on cravings/appetite and PFC.


Post a Comment

<< Home

eXTReMe Tracker