Predictive validity of three eating measures

The following is a guest post by Nicole Giuliani.

Food researchers (such as myself) study topics such as how often people crave their favorite foods or how well they are able to control those desires. We use a variety of laboratory measurement instruments as a proxy for how food craving and control work in the real world. However, it is not yet known how well many of those instruments—including the most common ones—predict actual food consumption out in the real world. Furthermore, we do not know how well the measurements we take in the laboratory predict real world eating for people who are on a diet versus people who are eating without restriction. Three common types of measurements are: people’s reports of how much they generally tend to desire certain foods and to regulate those desires (which we call “stable” tendencies to desire and regulate desire); people’s reports of the desirability of those foods at the moment that they view them or try to control their desire for them (“momentary”), and how people’s brain activity changes when they look at pictures of their desired foods or try to control their desire for those foods (“neural”).

My colleagues Elliot Berkman, Traci Mann, Janet Tomiyama and I recently published a study in the journal Psychosomatic Medicine in which we directly compared how well each of these three types of laboratory measurements predicted real world consumption of a favored tasty but unhealthy food. We brought a group of healthy people into the lab, asked them what their favorite unhealthy food was, and then gathered stable, momentary, and neural measurements of their craving for that food and their ability and tendency to regulate that craving. At the very end of the lab session, we split our participants into two random groups: one in which participants would simply record how much of their favorite unhealthy food they ate each day across the following two weeks (“monitor” group), and another in which participants tried not to eat their food and also recorded how much of it they ate (“restrict” group).

Graphs representing the relationship between Stable, Momentary, and Neural Reactivity (a) and Regulation (b) and Predicted Total Target Food Servings Consumed. Participants instructed to restrict target food consumption are shown in gray, participants instructed to simply monitor consumption (control) are shown in black. The x-axis scales represent the degree of reactivity (top row) or regulation (bottom row) in standard units, and reflect a composite of the measures based on independent components analyses as described in the text; y-axis scales are the predicted total number of target food servings consumed across the two-week sampling period. * ME p < 0.05; ^ interaction p < 0.05

Graphs representing the relationship between Stable, Momentary, and Neural Reactivity (a) and Regulation (b) and Predicted Total Target Food Servings Consumed. Participants instructed to restrict target food consumption are shown in gray, participants instructed to simply monitor consumption (control) are shown in black. The x-axis scales represent the degree of reactivity (top row) or regulation (bottom row) in standard units, and reflect a composite of the measures based on independent components analyses as described in the text; y-axis scales are the predicted total number of target food servings consumed across the two-week sampling period. * ME p < 0.05; ^ interaction p < 0.05

This experiment revealed some interesting differences in how well the instruments predicted actual food consumption in the two groups. First, higher scores on the momentary and neural measurements of food craving predicted more food consumption only for the monitor group, and had no relationship with eating in the restriction group. In other words, people who said they craved foods when exposed to them (and whose brains reacted strongly to those foods) generally ate more, but only when they weren’t trying to diet. Second, greater scores on the stable and momentary measurements of food regulation predicted less eating in the monitor group, whereas greater scores on the stable and neural measurements of food regulation predicted more consumption for the restriction group. That is, people who say they are “good regulators” do indeed eat less—but only when they aren’t trying to. When they try to diet, they actually eat more! This set of results is important because it suggests that food researchers should chose their measurements carefully, in part based on whether or not their participants will be restricting their food intake.

Citation info: Giuliani, N.R., Tomiyama, A.J., Mann, T., & Berkman, E.T.  (in press). Prediction of daily food intake as a function of measurement modality and restriction status. Psychosomatic Medicine.

Posted in News | Tagged , , , , , | Leave a comment

New grant from the National Institute on Aging for inhibitory control training

Our latest translational neuroscience project tests whether a computerized training program can reverse the effects that early adverse experiences have on inhibitory control. The project is funded by the National Institute on Aging and will last two years.

Early adversity (EA) in humans is a major contributing factor to a range of deleterious physical and mental health outcomes extending through adulthood such as depression and anxiety, obesity and heart disease, and premature death. In addition to detracting significantly from individual well-being and quality of life, these conditios also consume considerable resources from federal, state, and community organizations. The mechanisms through which EA exerts its effects on these outcomes are increasingly well understood, and include neurocognitive pathways related to executive function. An intervention that can successfully target, engage with, and alter the functioning of one or more of these mechanisms would be a promising way of mitigating the impact of EA on deleterious outcomes later in life. The proposed research focuses on one such pathway-deficits in inhibitory control (IC)-and tests the feasibility and efficacy of an intervention to increase functioning in that pathway in a sample of individuals who experienced EA. The intervention is grounded in a neurally informed model of change that specifies deficits in IC as an underlying causal factor common to several health-risking behaviors (HRBs). These IC deficits emerge during development as a result of a range of EA, and, critically, can be remediated in mid-life through targeted intervention. Research from our laboratory has validated an intervention that can increase IC performance and alter its underlying neural systems in young adults (Berkman, Kahn, & Merchant, 2014). The next step in this program of research, proposed here, is to test the efficacy of that intervention in a sample of mid-life individuals who have experienced EA and the extent to which our intervention generalizes to HRBs that are prevalent in that sample. The first Aim is to test whether the intervention alters the IC system in tasks both similar to and dissimilar from the training task in terms of both behavioral performance and neural functioning. The second Aim is to test whether alterations in the functioning of the underlying neural systems mediate the effect of the intervention on performance and The two Aims will be accomplished within the context of a single RCT with two arms (IC training vs. active control) and pre-post measurements of IC performance, IC neural systems, and HRBs. All participants (N = 110) come to the lab for an initial assessment of behavioral / neural measures of IC and HRBs, among other measures. Then, participants are randomly assigned to receive a Person-Centered Inhibitory Control (PeCIC) training or active control training, every other day for 3-4 weeks. The PeCIC systematically pairs IC engagement with alcohol, tobacco, and/or energy-dense food cues, depending on each participant’s reports of disinhibited behavior in those domains. The active control task uses personalized cues and response time tasks but does not involve IC. Finally, participants return to the laboratory for an endpoint assessment where all baseline measures are repeated. The two Aims will be robustly tested in a series of analyses comparing the behavioral and neural change from pre- to post-intervention between the groups disinhibition-related HRBs.

Posted in News | Comments Off on New grant from the National Institute on Aging for inhibitory control training

New grant from the National Cancer Institute to study food craving regulation

The lab has received funding from the National Cancer Institute (NCI) for a project entitled, “Reducing craving for cancer-promoting foods via cognitive self-regulation”. You can read more about the grant here.

The grant will last for two years beginning this fall.

Eating energy-dense foods when one is not hungry is a contributor to overweight and obesity, which are risk factors for a range of cancers. Excessive eating of a subset of these foods, such as red meat or foods with a high glycemic index, is an additional risk factor for cancer, separate from overweight and obesity. We refer to foods that are linked to cancer through either or both routes as cancer-promoting foods. The goal of this project is to reduce cancer risk by improving cognitive self-regulation of cravings for cancer-promoting foods. We focus on craving of cancer-promoting foods as one proximal determinant of their consumption. Craving consists of a subjective sense of wanting to eat a food, a motivation to seek out the food, and recurrent or intrusive thoughts related to the food. Considerable research shows that craving is a strong predictor of eating, even in the absence of hunger. Thus, enhancing a simple, low-cost and easily disseminated tool to reduce craving for cancer-promoting foods would advance cancer prevention and related research. Studies from affective science and social neuroscience have identified cognitive self-regulation strategies that are effective in reducing craving and their associated neural systems. This work has focused mostly on craving for other appetitive stimuli (e.g., drug cues), and has only begun to study regulation of food craving. Recent results from our laboratory validated four strategies that are effective in reducing cravings for energy-dense foods. This work relies upon self-reports of craving, which provide an empirical starting point but do not demonstrate the validity of the strategies on their own. Thus, the goals of the proposed project are to provide additional support for the effectiveness of cognitive self-regulation of food cravings using other measures beyond self-report, and to validate a theoretically grounded means to further increase the efficacy of those strategies-strategy choice. These goals will be accomplished in the context of a single study with two sessions. First, participants will be randomly assigned to choose their regulation strategy or to have one selected for them. In Session 1, their self-reported cravings and neural responses will be recorded while they alternately view images of energy- dense foods and regulate their responses to those foods. These data will be used to examine the effects of food regulation on neural activation and self-reports of craving, and to compare the effect of strategy choice on those measures. In Session 2, participants will visit our behavioral laboratory three days following Session 1 for a session in which their actual intake of energy-dense food will be measured in a naturalistic and unobtrusive manner. The difference in energy-dense food intake between the two groups will provide a behavioral measure of the efficacy of strategy choice on eating. Also, brain activity during food cue reactivity and regulation from Session 1 will be used to predict intake during Session 2. Psychological theory and previous neuroscience data suggest that neural activity, particularly in the medial prefrontal cortex, might explain variance in energy-dense food intake above and beyond self-report, and might mediate the effect of strategy choice on intake.

Posted in News | Comments Off on New grant from the National Cancer Institute to study food craving regulation

When does feedback increase or decrease motivation?

See Jordan Miller-Ziegler‘s thoughts on this question over at the lab blog on Psychology Today, The Motivated Brain.

Posted in News | Comments Off on When does feedback increase or decrease motivation?

Did Facebook really manipulate emotion?

Elliot is quoted in a piece by Maria Konnikova over at the New Yorker about the somewhat controversial Facebook study on emotional contagion.

Posted in News | Tagged , , | Leave a comment

Me, My Brain, and I

Check out Elliot’s latest blog post on the Motivated Brain at Psychology Today where he describes a recent symposium on self and self-functions from the 2014 Association for Psychological Science meeting.

“Social psychologists have made recently breakthroughs in understanding the self and its functions using neuroimaging. I discuss some of these discoveries, including the positive bias in self-perception, an apparent purpose for consciousness, and one surprising source of self-regulation. It turns out our brains contain some interesting information about ourselves!”

Posted in News | Tagged , , , , | Leave a comment

Comparison of text messaging and pencil-and-paper for ecological momentary assessment of food craving and intake

It is becoming increasingly common for researchers to use text messaging to measure everyday thoughts, feelings, and behaviors because of its low cost and people’s generally high levels of familiarity and comfort with the technology. Another key advantage of text messaging is its ability to measure time-sensitive variables such as cravings (for food or other substances) and hunger. Despite these advantages, there have been very few studies that systematically compare text messaging to alternative ways to measure those variables such as paper-and-pencil diaries. As such, text messaging remains a promising though (as yet) unevaluated technique to measure everyday experiences.

Because of our curiosity about eating behavior (among other everyday goals), we were particularly interested to confirm our hunch that text messaging would be a superior way to measure eating-related variables (intake, craving, etc) compared to older techniques for accomplishing the same task. In particular, we focused on paper-and-pencil diaries, which are quite common in the literature both historically and presently. In this study, we assigned participants to report on their food craving, food intake, and general hunger four times each day for two weeks using either text messaging (with our partner mProve Health) or using traditional paper-and-pencil diaries (see example below).

Paper and Pencil Diary

An example of a paper and pencil diary, one traditional way to measure everyday experience in vivo

Our main dependent measures were response rate, or how many responses were registered out of the 56 opportunities, and response latency, or how quickly after the targeted time did the participant respond. For example, a participant who responded 3 out of 4 times on a day would have a response rate of 75%, and a response that came at 2:02pm for a targeted time of 2:00pm would have a response latency of 2 minutes.

Our results were decisive: participants using the paper-and-pencil method had a response rate of 70%–pretty good, all things considered–but the response rate for participants in the text messaging condition was 96%. The response latencies were also shorter in the text messaging group (about 29 minutes after the targeted time on average) compared to the paper-and-pencil group (79 minutes after the targeted time on average). Interestingly, the absolute correlation of the response rate and latency with body mass index (BMI) was higher in the paper-and-pencil group than the text messaging group, suggesting that some of the subject-to-subject variability in responses in paper-and-pencil studies might be systematic (as a function of BMI in this case), but that text messaging reduces this form of variability.

Response latencies by group

Latencies in the paper-and-pencil group (black) and the text messaging group (white). The mean latency with respect to the targeted response time was far shorter in the text group (29 min) than it was in the paper-and-pencil group (79 min).

What we make of these findings is simple: text messaging makes it easier for participants to respond, so they respond more often and in a more timely way. We’re really happy about these findings because they establish text messaging as a valid way to help researchers get larger amounts of high-quality data on time-sensitive experiences such as food cravings.

My co-authors on this study were Nicole Giuliani and Alicia Pruitt.

Citation info: Berkman, E.T., Giuliani, N.R., & Pruitt, A.K. (2014). Comparison of text messaging and pencil-and-paper for ecological momentary assessment of food craving and intake. Appetite, 81, 131-137. [pdf]

Posted in News | Tagged , , | Comments Off on Comparison of text messaging and pencil-and-paper for ecological momentary assessment of food craving and intake

Lab manager position in the UO SAN/DSN labs!

The Developmental and Social Affective Neuroscience Labs, under the supervision of Profs. Jennifer Pfeifer and Elliot Berkman in the Department of Psychology at the University of Oregon, are currently seeking a lab manager to support studies of motivation, self-regulation, self-evaluation and decision-making, to begin mid-Summer 2014 (exact dates flexible). The position requires a commitment of at least one year.

Our research uses a combination of fMRI, behavioral, and longitudinal methods with adolescents and adults. Responsibilities will include assisting with the coordination and testing of human subjects, acquiring and analyzing fMRI data, programming stimuli, web design, managing IRB submissions and protocols, and coordinating day-to-day lab administration. The ideal candidate would also be excited to make original, creative contributions to the laboratory’s research.

Candidates must have a bachelor’s degree in psychology, neuroscience, or a related field. Some experience with fMRI and computer programming (e.g., MATLAB, R, Python) is preferred but not a prerequisite; however, candidates who are not yet proficient with a programming language must feel comfortable learning technical aspects of programming and data analysis. Experience working with a range of populations including adolescents, high-risk individuals, and adult community samples – or a desire to learn about doing so – is also highly valued. Candidates must be highly motivated, organized, and have excellent interpersonal skills.

For more information about the labs, visit (developmental side) and (social side). Please apply online ( regarding posting number 140506 with a cover letter indicating interest in this specific position, a current resume, references, and a list of relevant courses completed. Also, please be sure to mention Drs. Pfeifer and Berkman in your cover letter.

Posted in News | Tagged , , , , | Leave a comment

New paper from SAN lab members Will and Lauren!

Will and Lauren are authors on a new paper in the journal Social Cognitive and Affective Neuroscience about vmPFC activation during personal relevance and self-similarity judgments. Congrats, Will and Lauren (and Jenn)!

Moore, W. E., Merchant, J. S., Kahn, L. E., & Pfeifer, J. H. (2013). ‘Like me?’: Ventromedial prefrontal cortex is sensitive to both personal relevance and self-similarity during social comparisons. Social Cognitive and Affective Neuroscience, 9, 421-426.

Posted in News | Tagged , , | Leave a comment

Meditation, Gratitude, and Endogenous Opioids

Congrats to Lisa May for receiving a Varela Award from the Mind & Life Institute! This award complements her Dissertation Research Award from the Greater Good Science Center to support her work on the neurobiological mechanisms of pain relief.

Lisa writes about the funded project:

The goal of this project is to explore how meditation and gratitude affect pain. Quite a few scientific studies have documented that people who have a regular habit of meditating often feel less pain or perceive pain differently than people who don’t meditate. And there are good reasons to think that gratitude might cause pain relief, too. People who live with chronic pain often report that cultivating gratitude helps them deal with pain. Also, feeling pleasure, feeling like you’ve received something of value, or feeling motivated can cause pain relief, and gratitude incorporates pleasure, perceived value, and motivation!

What’s less well understood is how meditation and gratitude cause pain relief. One possibility is that gratitude and or meditation cause the release of endogenous opioids in the brain, leading to pain relief in exactly the same way morphine does. Other mental processes such as beliefs and expectations can release endogenous opioids, so it’s plausible that meditation or gratitude might. This study will test this by giving people an opioid antagonist drug called Naloxone. If Naloxone makes people’s meditation or gratitude-induced pain relief go away, that will be evidence that meditation and/or gratitude cause pain relief by activating opioid receptors.

Posted in News | Tagged , , , | 1 Comment