February 2014  << Back  

  Can't open the PDF? Click here for help.



Neurocognitive functioning in compulsive buying disorder

Katherine L. Derbyshire, BS

Department of Psychiatry and Behavioral Neuroscience, University of Chicago, Chicago, Illinois, USA

Samuel R. Chamberlain, MD, PhD

Department of Psychiatry, University of Cambridge, Cambridge, United Kingdom, Cambridge and Peterborough NHS Foundation Trust (CPFT), Cambridge, United Kingdom

Brian L. Odlaug, MPH

Department of Public Health, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

Liana R. N. Schreiber, BA

Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, Minnesota, USA

Jon E. Grant, JD, MD, MPH

Department of Psychiatry and Behavioral Neuroscience, University of Chicago, Chicago, Illinois, USA

BACKGROUND: Compulsive buying (CB) is a fairly common behavioral problem estimated to affect 5.8% of the population. Although previous research has examined the clinical characteristics of CB, little research has examined whether people with CB manifest cognitive deficits.

METHODS: Twenty-three non–treatment-seeking compulsive buyers (mean age, 22.3±3.5; 60.9% female) and 23 age- and sex-matched healthy controls (mean age, 21.1±3.4, 60.9% female) underwent neurocognitive assessment. We predicted that the following cognitive domains would be impaired in CB: spatial working memory (Spatial Working Memory test), response inhibition (Stop-Signal Task), cognitive flexibility (Intra-Extra Dimensional Set Shift task), and decision making (Cambridge Gambling Task).

RESULTS: Compared with controls, individuals with CB exhibited significant impairments in response inhibition (P=.043), risk adjustment during decision making (P=.010), and spatial working memory (P=.041 total errors; P=.044 strategy scores). Deficits were of large effect size (Cohen’s d, 0.6 to 1.05).

CONCLUSIONS: These pilot data suggest that individuals with CB experience problems in several distinct cognitive domains, supporting a likely neurobiological overlap between CB and other putative behavioral and substance addictions. These findings may have implications for shared treatment approaches as well as how we currently classify and understand CB.

KEYWORDS: compulsive, buying, neurocognition



Compulsive buying (CB) is characterized by preoccupation with buying/shopping and frequent buying that causes substantial psychological, social, occupational, and/or financial impairment.1 CB has an estimated prevalence of 5.8% to 6.9% in the US and German population-based samples (mean age, 39.7±15.72; mean age, 44.3±14.93) although lower rates (1.9%) have been found among college students (mean age, 20.0±1.3 years4). Buying episodes are typically preceded by feelings of low mood, tension, or boredom, with a sense of relief or pleasure typically occurring during shopping episodes. Feelings of remorse, guilt, and shame often follow shopping episodes.1 The purchased items are typically hidden, hoarded, given away, or discarded. Individuals with CB frequently accumulate large amounts of debt and report being unable to control the behavior despite the negative consequences.5 CB is associated with high rates of psychiatric comorbidity, in particular, depression, anxiety, and compulsive hoarding.6 Studies of clinical samples suggest that females are predominantly affected by CB and experience financial distress, marital problems, issues within the family, legal problems, and have a great amount of personal distress.2,5,6

Despite a fairly high population prevalence and associated morbidity, virtually no research has examined cognitive functioning in individuals with CB. The only published study to date examined 26 individuals with CB vs 22 healthy controls on a variety of measures, including the Wisconsin Card Sorting Test (WCST), Iowa Gambling Task (IGT), and verbal learning paradigms.7 Contrary to the authors’ hypotheses, people with CB did not exhibit significant cognitive dysfunction vs controls, although performance was numerically worse on several measures. Contrary to expectations, people with CB performed significantly better than controls on a single measure relating to picture completion.

In the relative absence of cognitive studies in CB, insights into the cognitive problems likely to be associated with CB can be gleaned by considering cognitive findings in related conditions. CB shares phenomenological similarities and comorbid overlap with behavioral addictions such as gambling disorder.8 Gambling disorder has been associated with impairments on tests of frontal lobe integrity, including tests of cognitive flexibility, response inhibition, executive planning,9 and decision making (eg, the Cambridge Gambling Task [CGT]), tasks that are dependent on the integrity of the orbitofrontal cortices and other affective-related neural circuitry.10 Neural and neurochemical substrates of these computerized paradigms have in many cases been explored across disorders and, in some cases, across species.11

The aims of the present study are to investigate cognitive function in young adults with current CB compared with controls using a range of validated translational computerized paradigms known to be sensitive to frontostriatal function. We hypothesized that individuals with CB would exhibit impairments on aspects of decision making, impulse control, and working memory, consistent with underlying dysregulation of frontostriatal circuitry and overlapping neurobiology with other behavioral and substance addictions.



Four hundred and twenty-six subjects were recruited via media advertisements for an observational longitudinal study examining impulsivity in young adults, in which they completed several assessments, including a neurocognitive battery (Cambridge Neuropsychological Test Automated Battery [CANTAB]). All subjects completed the self-reported Minnesota Impulse Disorders Interview (MIDI),6,12 and 23 individuals met proposed criteria for current CB.5 Individuals were enrolled based on their ability to understand/undertake the procedures and to provide voluntary written consent. Twenty-three age- and sex-matched controls (ie, with no current [past 12 months] Axis I psychiatric disorders) were randomly selected from the population being sampled. The study procedures were carried out in accordance with the ethical standards established by the Declaration of Helsinki. The institutional review board of the University of Minnesota approved the study and the consent procedures. After all study procedures were explained to the participants, voluntary written informed consent was obtained. Subjects were recruited between October 22, 2009 and April 10, 2012.

CB was diagnosed using the MIDI, which is based on criteria proposed by McElroy and colleagues: 1) preoccupation with buying (characterized by either an irresistible, intrusive, and/or senseless preoccupation with buying or buying more than one can afford; buying unneeded items; or shopping for longer durations of time than originally intended); and 2) the preoccupation with buying results in marked distress, interferes with social or occupational functioning, and causes financial problems.5 Subjects underwent psychiatric examination using the Mini International Neuropsychiatric Interview (MINI).13 Subjects also completed a brief 6-question screen of an attention-deficit/hyperactivity disorder (ADHD) scale, which was designed to be a short, convenient screening form that uses questions based on DSM-IV criteria for ADHD.14

Cognitive measures

Participants completed the following selected cognitive paradigms from the CANTAB (CANTAB Eclipse, version 3, Cambridge Cognition Ltd) in a fixed order (the order of administration was the same as the order of the described tasks below). Brief descriptions of each task are given; the reader is referred to the references for fuller descriptions and background on each paradigm.

Intra-Extradimensional Set Shift (IDED) task.15 The IDED task includes aspects of rule learning and behavioral flexibility. On each trial, volunteers are presented with 2 stimuli and asked to work out an underlying rule about which picture is “correct” based on feedback (“wrong” or “correct” presented after each selection). The primary outcome measure is the total number of errors made on the task, adjusting for stages not attempted (total errors adjusted). Where this measure shows a significant difference between groups of interest, subscores for different stages of the task can be considered individually, to further explore the nature of the impairment.

Stop Signal Task (SST).16 The SST is a test of response inhibition in which subjects respond to a series of directional arrows appearing one at a time on-screen (for a left arrow, they press a left button and vice versa). On a subset of trials, an auditory “stop signal” occurs, indicating to participants that they should try to suppress their motor response for that given trial. Inhibition on this and related tasks is dependent on a right-lateralized neural network, including the inferior frontal gyrus. This task estimates the time taken by each volunteer’s brain to suppress an already triggered command (the stop-signal reaction time [SSRT]). Longer SSRTs correspond to worse inhibitory control. The task also records median reaction time (RT) for “go” trials, which is a measure of general response latencies.

Cambridge Gambling Task (CGT).17 The CGT measures aspects of decision making and has been shown to be sensitive to frontal lesions. On each trial, participants are presented with a mix of red and blue boxes on-screen and are told that the computer has hidden a “token” behind one of them; they have to choose a) what color of box they believe the token is hidden behind (red or blue), and b) the number of accumulated points they want to gamble on having made the correct color choice. The proportion of red to blue boxes (box ratio) is varied over the task pseudorandomly to assess the influence of statistical risk on decision making. The key outcome measures for this task are the proportion of rational decisions made (choosing the most logical color—the one in the majority on the screen), the proportion of points gambled, and the extent of risk adjustment (a measure of the extent to which betting behavior, ie, amount gambled, is varied by subjects as a function of statistical risk).

Spatial Working Memory (SWM).18 On the SWM test (8-box version), participants attempt to locate tokens hidden underneath boxes on-screen and try to avoid returning to boxes that previously yielded such tokens. The key outcome measures include the “total number of errors” (inappropriately returning to boxes that previously yielded tokens) and “strategy score” (lower score equates to superior strategy use).

Data analysis

To compare demographic, clinical, and cognitive characteristics between the groups, one-way analysis of variance (ANOVA) tests were used, or nonparametric equivalents where appropriate. This being a pilot study, significance was defined as P < .05, uncorrected. Where significant differences between the study groups were identified on given measures, effect sizes were reported (Cohen’s d). The data were analyzed using SPSS, version 19 (IBM).


Out of our total sample of 426 individuals, 344 completed the MINI and a total of 23 compulsive buyers (6.7%; mean age, 22.3±3.5; 60.9% female; 78.3% white) were identified. The 23 healthy controls did not differ significantly from compulsive buyers in terms of age and sex (mean age, 21.1±3.4; 60.9% female; 82.2% white) or on any other demographic variables (TABLE 1).


Demographic variables between groups

Characteristic Compulsive buyers Healthy controls P
Mean (SD) 22.3±3.5 21.1±3.4 .197
Female 14 (60.9%) 14 (60.9%) 1.000a
White 18 (78.3%) 19 (82.2%) .718a
Relationship status
Single 23 (100%) 22 (95.7%) .323a
Educational level
Less than high school 1 (4.35%) 0 .323a
High school graduate 1 (4.35%) 0 .323a
Some college 15 (65.2%) 18 (78.3%) .337a
College graduate 4 (17.4%) 2 (8.7%) .393a
College plus 2 (8.7%) 3 (13.0%) .645a
aDenotes a chi-square result.
Compulsive buying sample

Of the 23 subjects with CB, all reported financial problems secondary to their shopping behavior. In addition, 21.8% (n=5) reported social or family problems resulting from buying (21.8%), and 13.0% (n=3) reported that their shopping had resulted in relationship problems.

Eleven (47.8%) of the compulsive buyers met criteria for at least one additional current psychiatric diagnosis, with the most common being major depressive disorder (n=3; 13.0%) and substance use disorder (n=4; 17.4%). The CB and control groups did not differ from each other on ADHD scale total scores (CB mean=10.5, control mean=13.0; P=.284).

Neurocognitive performance

Performances of the groups on the cognitive measures of interest are summarized in TABLE 2. Subjects with CB made significantly more errors than controls on the spatial working memory task (P=.041), and this was accompanied by significantly worse strategy scores (as higher strategy scores indicate worse performance; P=.044). Individuals with CB exhibited poorer decision making on the CGT specifically in terms of making less risk adjustment (P=.010) in comparison with the controls. Finally, individuals with CB demonstrated greater impulsivity on the SST (P=.043). No significant differences were noted on the IDED task (P=.922). Secondary analyses found no significant differences on cognitive measures between CB subjects with and without psychiatric comorbidity (all P > .05) (TABLE 3).


Cognitive task comparison between groups

Task Compulsive buyers (mean, SD) Non-compulsive buyers (mean, SD) P d
IDED total errors (adjusted) 26.8±21.9 27.4±19.7 .922  
SST SSRT 220.6±115.5 162.4±48.1 .043 .658
SST median correct RT on go trials 414.5±116.8 465.9±122.8 .401  
CGT overall proportion bet .59±.13 .46±.16 .076  
CGT quality of decision making .94±.11 .95±.06 .771  
CGT risk adjustment 1.2±1.1 2.0±.85 .010 -.814
SWM strategy 30.5±7.2 24.2±4.5 .044 1.049
SWM total errors 19.4±17.8 6.3±6.5 .041 .977
CGT: Cambridge Gambling Task; d: Cohen’s d; IDED: Intra-Extradimensional Set Shift task; RT: reaction time; SSRT: stop-signal reaction time; SST: Stop-Signal Task; SWM: Spatial Working Memory.


Cognitive task comparison among the CB group with and without psychiatric comorbidity

Task With psychiatric comorbidity (mean, SD) Without psychiatric comorbidity (mean, SD) P
IDED total errors (adjusted) 26.8±22.3 26.8±22.6 .994
SST SSRT 222.1±73.3 219.2±147.6 .955
SST median correct RT on go trials 498.4±240.0 553.9±396.3 .692
CGT overall proportion bet .61±.16 .58±.11 .579
CGT quality of decision making .91±.14 .97±.05 .229
CGT risk adjustment 1.0±1.1 1.4±1.1 .482
SWM strategy 32.2±6.8 28.9±7.4 .286
SWM total errors 23.9±19.8 15.3±15.5 .259
CB: compulsive buying; CGT: Cambridge Gambling Task; IDED: Intra-Extradimensional Set Shift task; RT: reaction time; SSRT: stop-signal reaction time; SST: Stop-Signal Task; SWM: Spatial Working Memory.


The purpose of this study was to examine the cognitive performance of individuals with CB as compared with healthy volunteers. We found that CB was associated with inferior scores in comparison with the control group on spatial working memory impairment (SWM) and motor impulse control (SST), and were found to have select problems with adapting behavior in light of risk on the gambling task (CGT). These results appeared to be unrelated to ADHD symptoms (which previous research has suggested may account for cognitive deficits in CB7), since our sample did not differ from controls on ADHD symptom scores.

In this study, individuals with CB exhibited greater problems in SWM, both in terms of the number of errors made, and in terms of strategy scores—deficits that also have been reported in a meta-analysis of ADHD studies, studies of young binge and non-binge social drinkers, and adolescents with alcohol and marijuana dependence.11,19-21 The function of the working memory may be of special interest in CB, as it influences self-regulation. In natural reward (food, sex), individuals with low working memory capacity show more automatic behavior than individuals with high working memory capacity.22 This supports the crucial role of working memory in self-regulating behavior, which may reflect the lack of control over shopping seen in CB.

The SST, a task that measures the ability to suppress a prepotent motor response, differed significantly between those with and without CB. The SST is dependent on the right inferior frontal gyrus and associated regions,16 which has been found to be impaired in conditions such as ADHD and gambling disorder.23-25 The motor impulsivity seen on this task mirrors the clinical presentation of individuals with CB. Individuals with CB often report that once they start buying, they cannot control their behavior even if they are aware of the possible consequences.1

Decision making and risk taking (as reflected by the CGT) were also poorer in the CB subjects. The CB group exhibited significantly less risk adjustment than controls on the task, indicating that, over the course of the task, they were insufficiently sensitive to changes in statistical risk and did not appropriately adjust the amount of points they gambled, depending on this changing risk. Individuals with CB tend to persist in spending despite rising credit card debt and other financial consequences.1 This finding seems to reflect the clinical symptoms of the disorder. Impaired risk-adjustment has been observed in people with insular cortex lesions,26 and also has been associated with white matter abnormalities in the thalamus and dorsal striatum in individuals with traumatic brain injury.27 The current findings may implicate insular cortex and relative insensitivity to aversive outcomes in the pathophysiology of CB.

The CB group was not found to be significantly different when compared on all cognitive domains considered (no significant differences were detected on the IDED task in terms of total errors on the task). The IDED quantifies aspects of mental flexibility, so in that regard, the CB group did not appear to have a disadvantage in learning rules and then adapting to changes in the task.

Overall, deficiencies in the CB group spanned a range of cognitive domains and included spatial working memory, risk adjustment, and inhibition of behaviors. Research has suggested that impulse control disorders may be neurobiologically similar to addictions,28 which can also be seen here with similarities to substance use disorders,21,29-30 and/or to ADHD.31 Further examination of the relationship of CB to other disorders will be important to properly characterize the disorder and to develop treatment approaches.

This study has several limitations that should be considered. First, the assessment for CB (the MIDI) was self-reported and therefore may have over- or underdiagnosed CB. The MIDI has, however, been used in other self-report studies with good reliability and validity.4 Second, the sample was fairly homogenous with a high percentage of white individuals (78.3%), and therefore the results may not generalize to more racially or ethnically diverse populations. Third, approximately one-half of those who endorsed CB on the MINI had a concurrent psychiatric diagnosis, so it is difficult to tease apart how much influence these disorders had over the results. We found no differences, however, in cognitive measures between CB subjects with and without comorbidities. Larger sample sizes would be needed to more fully investigate the contributing influences of other conditions, but we believe our study represents a vital first step in this process, given the paucity of research on this topic. Finally, the sample size was relatively small, and this may have resulted in limited power to detect subtle cognitive inferiority in CB vs controls.


To our knowledge, this is the first study to examine cognition in CB using a range of well-validated computerized paradigms. With the high prevalence rating of this disorder, it is important to continue research to examine possible long-term cognitive consequences of CB or examine cognitive precursors for the development of CB. This study found that those with CB manifested spatial working memory impairment, impulse dyscontrol, and less risk adjustment than controls during decision making. Future cognitive testing, including functional brain imaging, may aid in the identification of other cognitive deficiencies and in characterizing neural abnormalities associated with these cognitive deficits.

DISCLOSURES: This research was supported in part by the National Center for Responsible Gaming and an American Recovery and Reinvestment Act Grant from the National Institute on Drug Abuse (NIDA) (1RC1DA028279-01) to Dr. Grant. Dr. Grant has received research grants from the National Institute of Mental Health, NIDA, National Center for Responsible Gaming, Forest, Transcept, Roche, and Psyadon Pharmaceuticals, and the University of South Florida. He receives yearly compensation from Springer Publishing for acting as Editor-in-Chief of the Journal of Gambling Studies and has received royalties from Oxford University Press, American Psychiatric Publishing, Inc., Norton Press, and McGraw Hill. Dr. Chamberlain has consulted for Cambridge Cognition, P1Vital, and Shire Pharmaceuticals, has received speaker honoraria from Eli Lilly and Company, and has received research grants from the Academy of Medical Sciences. Mr. Odlaug has received a research grant from the Trichotillomania Learning Center, is a consultant for Lundbeck Pharmaceuticals, and has received honoraria and royalties from Oxford University Press. Ms. Derbyshire and Ms. Schreiber report no conflicts of interest.


  1. Black D. A review of compulsive buying disorder. World Psychiatry. 2007;6:14–18.
  2. Koran L, Faber R, Aboujaoude E, et al. Estimated prevalence of compulsive buying behavior in the United States. Am J Psychiatry. 2006;163:1806–1812.
  3. Mueller A, Mitchell J, Crosby R, et al. Estimated prevalence of compulsive buying in Germany and its association with sociodemographic characteristics and depressive symptoms. Psychiatry Res. 2010;180:137–142.
  4. Odlaug B, Grant J. Impulse-control disorders in a college sample: results from the self-administered Minnesota Impulse Disorders Interview (MIDI). Prim Care Companion J Clin Psychiatry. 2010;12(2):e1–e5.
  5. McElroy S, Keck Jr, Pope H Jr, et al. Compulsive buying: a report of 20 cases. J Clin Psychiatry. 1994;55:242–248.
  6. Christenson G, Faber R, Dezwaan M. Compulsive buying: descriptive characteristics and psychiatric comorbidity. J Clin Psychiatry. 1994;55:5–11.
  7. Black D, Shaw M, McCormick B, et al. Neuropsychological performance, impulsivity, ADHD symptoms, and novelty seeking in compulsive buying disorder. Psychiatry Res. 2012;2-3:581–587.
  8. Grant J, Brewer J, Potenza M. The neurobiology of substance and behavioral addictions. CNS Spectrums. 2006;11:924–930.
  9. Goudriaan A, Oosterlaan J, De Beurs E, et al. Pathological gambling: a comprehensive review of biobehavioral findings. Neurosci Biobehav Rev. 2004; 28:123–141.
  10. Clark L. Decision-making during gambling: an integration of cognitive and psychobiological approaches. Philos Trans R Soc Lond B Biol Sci. 2010;365:319–330.
  11. Chamberlain S, Robbins T, Winder-Rhodes S, et al. Translational approaches to frontostriatal dysfunction in attention-deficit/hyperactivity disorder using a computerized neuropsychological battery. Biol Psychiatry. 2011;69:1192–1203.
  12. Grant JE. The Minnesota Impulse Disorders Interview. In: Grant JE. Impulse control disorders: a clinician’s guide to understanding and treating behavioral addictions. New York, NY: W.W. Norton & Company; 2008;187–194.
  13. Sheehan D, Lecrubier Y, Sheehan K, et al. The Mini-International Neuropsychiatric Interview (MINI): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998;59:22–33.
  14. Kessler R, Adler L, Gruber M, et al. Validity of the World Health Organization Adult ADHD Self-Report Scale (ASRS) Screener in a representative sample of health plan members. Int J Methods Psych Res. 2007;16:52–65.
  15. Owen A, Roberts A, Polkey C, et al. Extra-dimensional versus intra-dimensional set shifting performance following frontal lobe excisions, temporal lobe excisions or amygdalo-hippocampectomy in man. Neuropsychologia. 1991;29:993–1006.
  16. Aron A. The neural basis of inhibition in cognitive control. Neuroscientist. 2007;13:214–228.
  17. Rogers R, Owen A, Middleton H, et al. Choosing between small, likely rewards and large, unlikely rewards activates inferior and orbital prefrontal cortex. J Neurosci. 1999;19:9029–9038.
  18. Owen A, Downes J, Sahakian B, et al. Planning and spatial working memory following frontal lobe lesions in man. Neuropsychologia. 1990;28:1021–1034.
  19. Weissenborn R, Duka T. Acute alcohol effects on cognitive function in social drinkers: their relationship to drinking habits. Psychopharmacol. 2003;165:306–312.
  20. Townshend J, Duka T. Binge drinking cognitive performance and mood in a population of young social drinkers. Alcohol Clin Exp Res. 2005;29:317–325.
  21. Schweinsburg A, Schweinsburg B, Cheung E, et al. fMRI response to spatial working memory in adolescents with comorbid marijuana and alcohol use disorders. Drug Alcohol Depend. 2005;79:201–210.
  22. Hofmann W, Gschwendner T, Friese M, et al. Working memory capacity and self-regulatory behavior: toward an individual differences perspective on behavior determination by automatic versus controlled processes. J Pers Soc Psychol. 2008;95:962–977.
  23. Chamberlain S, Sahakian B. The neuropsychiatry of impulsivity. Curr Opin Psychiatry. 2007;20:255–261.
  24. Odlaug B, Chamberlain S, Kim S, et al. A neurocognitive comparison of cognitive flexibility and response inhibition in gamblers with varying degrees of clinical severity. Psychol Med. 2011;41:2111–2119.
  25. Brevers D, Cleeremans A, Verbruggen F, et al. Impulsive action but not impulsive choice determines problem gambling severity. PloS One. 2012;7:e50647.
  26. Clark L, Bechara A, Damasio H, et al. Differential effects of insular and ventromedial prefrontal cortex lesions on risky decision-making. Brain. 2008;131:1311–1322.
  27. Newcombe V, Outtrim J, Chatfield D, et al. Parcellating the neuroanatomical basis of impaired decision-making in traumatic brain injury. Brain. 2011;134:759–768.
  28. Brewer J, Potenza M. The neurobiology and genetics of impulse control disorders: relationships to drug addictions. Biochem Pharmacol. 2008;75:63–75.
  29. Bechara A. Disturbances of emotion regulation after focal brain lesions. Int Rev Neurobiol. 2004;62:159–193.
  30. Rogers Rd, Everitt B, Baldacchino A, et al. Dissociable deficits in the decision-making cognition of chronic amphetamine abusers, opiate abusers, patients with focal damage to prefrontal cortex, and tryptophan-depleted normal volunteers: evidence for monoaminergic mechanisms. Neuropsychopharmacology. 1999;20:322–339.
  31. Barkley R, Brown T. Unrecognized attention-deficit/hyperactivity disorder in adults presenting with other psychiatric disorders. CNS Spectrums. 2008;13:977–984.

CORRESPONDENCE: Katherine L. Derbyshire, BS Department of Psychiatry and Behavioral Neuroscience University of Chicago 5841 S. Maryland Avenue, MC 3077 Chicago, IL 60637 USA E-MAIL: kderbyshire@uchicago.edu