Sunday, August 24, 2008

Borderline … feels like I'm goin' to lose my mind

No, the 80s Madonna song isn't really about Borderline Personality Disorder. But a new study in Science (King-Casas et al., 2008) makes me feel like I'm goin' to lose my mind. Or more precisely, makes me exhibit two of nine DSM-IV-TR criteria:
6. Affective instability due to a marked reactivity of mood (e.g., intense episodic dysphoria, irritability, or anxiety usually lasting a few hours and only rarely more than a few days).

8. Inappropriate anger or difficulty controlling anger (e.g., frequent displays of temper, constant anger, recurrent physical fights, getting mad over something small).
All right all right, a stigmatizing label is no laughing matter.

Roth and Fonagy (1996) defined BPD thusly:
The essential feature of this disorder is a pervasive pattern of instability of self-image, interpersonal relationships and mood. The person’s sense of identity is profoundly uncertain. Interpersonal relationships are unstable and intense, fluctuating between the extremes of idealisation and devaluation. There is often a terror of being alone, with great efforts made to avoid real or imagined abandonment. Affect is extremely unstable, with marked shifts from baseline mood to depression and anxiety usually lasting a few hours. Inappropriate anger and impulsive behaviour are common, and often this behaviour is self-harming. Suicidal threats and self-mutilation are common in more severe forms of this disorder.
So it's a very serious personality disorder. However, the diagnostic label is a controversial one, coming under fire from feminists (Shaw & Proctor, 2005) and from my favorite psychiatry blog, Shrink Rap. Here's why Dinah doesn't like the label:
Over The Border Line

...I'm going to talk about why I hate the term, why I rarely place it in writing, why I wish it would go away.

. . . [read the nine diagnostic criteria for BORDERLINE PERSONALITY DISORDER at Wikipedia]

I mean, okay, these symptoms cluster in some people, so why don't I like them? Here goes, with no particular rhyme or reason:
  • The diagnosis (unlike, say, Trichotillomania or Major Depression) is pejorative.
  • Clinicians are sloppy with the diagnosis and it's not uncommon for a doc to refer to a patient as "a borderline" as a defense--- the patient is difficult to deal with, he's angry or demanding--it's gotta be him, not the doc.
  • It's what clinicians label patients they don't like.
  • Treatment-wise, many docs avoid these patients and hope runs dry quickly. The prognostic implications are generally not great, these patient don't have rapid and dramatic improvements.
  • The diagnosis ends up being it's own endpoint, it doesn't leave room for alternate explanations and sometimes patients with Bipolar Disorder look a lot like patients with borderline personality disorder. Oh, while I'm there, patients with Borderline Personality Disorder often have co-morbid Bipolar Disorder (and hey, how about some substance abuse issues thrown in) and if the clinician can get focused on treating the Mood Disorder, sometimes the other noise fades into the background.
  • It doesn't seem to me that every patient who has these symptoms has them forever in an inflexible way. They come, they go, they change, they get better, they get worse.

With all those caveats in mind, let's look at the lucky 08/08/08 Science paper entitled "The Rupture and Repair of Cooperation in Borderline Personality Disorder." It examines how well individuals with BPD play along with others in an economic exchange game (see also King-Casas et al., 2005, the first of their three Science papers on the topic).

The key participants were 38 controls and 55 people with BPD.1 The vast majority of these participants were female (37 controls, 51 BPD). The two groups were also matched for age and verbal IQ. Each person was paired with another individual (all controls) to play the game, to form a total of 93 dyads.

Who can you trust?

Distrust of others is a cardinal feature of BPD. A history of childhood abuse and neglect is very common in this disorder -- leading to an early rupture of trust that is not easily repaired. A recent paper by Bhar, Brown, and Beck (2008) examined the factor structure of the Personality Beliefs Questionnaire (BPD subscale) and found three major factors: dependency, distrust, and the belief in acting preemptively to avoid threat. All three factors were associated with depression. Dependency and distrust were associated with hopelessness. Distrust, however, was the only factor that was significantly correlated with suicidal ideation.

Thus, assessing trust in BPD patients is a worthwhile endeavor. King-Casas et al. (2008) measured it using the trust game (reviewed in Camerer, 2003).

How does the "trust" game work? (see Read Montague give a brief explanation in Real Player). One player (the Investor) gives a sum of money to the other player (the Trustee). The investment triples, and the Trustee decides how much to give back to the Investor.

You can also read The Neurocritic's previous post on The Trust Game, which discussed the paper by Tomlin, Kayali, King-Casas, et al. (2006).2
I give the authors credit for developing the nifty "hyperscanning" methodology, which involves two subjects who interact with each other while their brains are scanned simultaneously (in adjoining or distant magnets). The task is illustrated below. Brain images are taken at three critical "reveal" points:

The comparison groups in the current experiment played the Trustee role in 10 rounds of the game. The results indicated that in the early rounds, Investors gave BPD and control Trustees similar amounts of money. In the latter 5 rounds of the game, however, there was a decline in the investments given to BPD Trustees but not to control Trustees, because the former weren't as generous with their returns overall. Specifically, the BPD subjects were less likely to use a "coaxing" strategy:
Healthy trustees are twice as likely as BPD trustees to coax when cooperation between players is low. Specifically, healthy trustees are more likely to make a large repayment (greater than or equal to investment amount) after having received a small investment (less than or equal to $5). Conversely, BPD trustees are more likely to make a small repayment (less than the investment amount) after receiving a small investment.

In general, players do not always behave rationally. To maintain higher investments in later rounds, however, coaxing is a beneficial strategy for the Trustee. Maybe levels of the natural "trust drug" oxytocin are low in BPD (ask Paul Zak, or read this PDF), or maybe they just aren't as good at mental arithmetic. Perhaps the BPD propensity to punish low investments could be viewed as analogous to what is seen in the Ultimatum Game (Sanfey et al., 2003) where players reject unfair offers (to their own detriment).

So why do I feel like I'm goin' to lose my mind?3 As suggested by a colleague, the present paper 1) makes liberal use of reverse inference and 2) reeks of fishing.4 In my view, the real trouble arises when the authors try to explain what bits of the brain might be implicated in the lack of trust shown by players with BPD. It's the insula! [and only the insula]. Why is that problematic? We shall return to that question in a moment.

In the study of Sanfey et al., unfair offers were associated with greater activity in bilateral anterior insula, dorsolateral prefrontal cortex, and anterior cingulate cortex, with the degree of insular activity related to the stinginess of the offer. A similar relationship was observed here in the controls, but not in the BPD patients. Taking a step back for a moment, we see differences between control and BPD participants (for the contrast low vs. high investment) in quite a number of places, as illustrated in the table below.

Table S3 (King-Casas et al., 2008). General linear model of Trustee brain regions with greater response to the revelation of small investments (less than or equal to $5) relative to large investments (greater than $10) in healthy Trustees relative to Trustees with BPD. (p less than .05, FDR corrected; min 5 voxels)

However, the within-group analysis in controls yielded a "small investment" effect only in bilateral anterior insula (12 voxels and 15 voxels, respectively, at p less than .10). The same analysis in the BPD group yielded absolutely no significant differences anywhere in the brain! BPD participants react to small stingy investments with differential behavior (by returning a very low percentage of the investment), yet there is no area in the brain telling them to do this. Perhaps something is going on in the delay period between the investment and repayment phases, but if so we don't find out. The authors' interpretation of the null brain effect is that BPD subjects have a social perception problem (as Montague explains here), and do not respond correctly to social norm violations.
The association of the insula with a representation of outcome variance suggests that the insula may encode the distribution of likely outcomes in social interactions; that is, responses in the anterior insula may indicate social norm violations within interpersonal contexts.
When dishing out small returns in the repayment phase, both control and BPD groups show the inverse correlation between insular activity and monetary amount. BPD, the apparent insensitivity of the insula only to offer level size, and not their own repayment, suggests two possibilities: (i) Monetary reward is not reinforcing to individuals with BPD; or, (ii) low offers are not perceived to be a violation of social norms to individuals with BPD.
They rule out the first possibility and go with the second.

But what about the role of dorsal striatum (King-Casas et al., 2005) and ventral tegmental area (Krueger et al., 2007) in the development of trust, or the importance of medial prefrontal cortex and theory of mind [see Mentalizing Mentalizing ]? We do not find out about these, either. Guess we'll have to wait for the next Science paper.

Finally, there's the wildly overblown quotes on the significance of the findings in the popular press:
Mind games reveal people who are 'blind' to trust

The finding ... could give psychiatrists a better diagnostic tool and a brain area to target with therapy or drugs when treating BDP, says project leader Read Montague, a neuroscientist at Baylor College of Medicine in Houston.

And this quote, which heralds the dawn of a golden new age of psychiatry:

Peter Bossaerts, a neuroeconomist at Caltech in Pasadena, says that since the games are scored and have a predicted outcome, they may help diagnose mental illnesses.

"This could really revolutionise the way people think about and treat psychiatric illnesses," he says.
Who can you trust?


1 Why those particular numbers? We don't know for sure, but it could be due to the fact that supplementary analyses were run with subgroups of the BPD patients to control for income level and medication status.

2 Oddly (or not), the present King-Casas paper did not even cite their earlier work (Tomlin et al., 2006). The authors appear to have taken the approach of "one brain region, one Science paper." Today it's the insula, in 2006 it was the anterior cingulate cortex, and in 2005 it was the dorsal striatum. How do these brain regions work together to produce the complex behavior in question? Are the three papers even consistent with each other? These issues are not important, apparently.

3 Other than the fact that I am not knowledgeable in behavioral game theory (see Camerer et al., 2003 for that, PDF).

4 They also reported significance (corrected using FDR procedures) at the p<.1 level. Why? This paper on Detecting signals in FMRI data using powerful FDR procedures (Pavlicova et al., 2008) recommends the standard α level (= .01 or .05).


Bhar SS, Brown GK, Beck AT. (2008). Dysfunctional beliefs and psychopathology in Borderline Personality Disorder. J Personal Disord. 22:165-77.

Camerer CF. (2003). Psychology and economics. Strategizing in the brain. Science 300:1673-5.

King-Casas B, Sharp C, Lomax-Bream L, Lohrenz T, Fonagy P, Montague PR. (2008). The Rupture and Repair of Cooperation in Borderline Personality Disorder. Science, 321(5890), 806-810. DOI: 10.1126/science.1156902

King-Casas B, Tomlin D, Anen C, Camerer CF, Quartz SR, Montague PR. (2005). Getting to know you: reputation and trust in a two-person economic exchange. Science 308:78-83.

Krueger F, McCabe K, Moll J, Kriegeskorte N, Zahn R, Strenziok M, Heinecke A, Grafman J. (2007). Neural correlates of trust. Proc Natl Acad Sci 104:20084-9.

Roth A, Fonagy P. (1996). What Works for Whom? A Critical Review of Psychotherapy Research. London and New York: Guilford.

Sanfey AG, Rilling JK, Aronson JA, Nystrom LE, Cohen JD. (2003). The neural basis of economic decision-making in the Ultimatum Game. Science 300:1755-8.

Shaw C, Proctor G. (2005). Women at the Margins: A Critique of the Diagnosis of Borderline Personality Disorder. Feminism and Psychology 15:483-490.

Tomlin D, Kayali MA, King-Casas B, Anen C, Camerer CF, Quartz SR, Montague PR. (2006). Agent-specific responses in the cingulate cortex during economic exchanges. Science 312:1047-50.

Borderline … feels like I'm goin' to lose my mind
You just keep on pushin' my love over the borderline
Borderline … feels like I'm goin' to lose my mind
You just keep on pushin' my love over the borderline (borderline)


Subscribe to Post Comments [Atom]


At August 29, 2008 4:09 PM, Anonymous Anonymous said...

Hi neurocritic. I need some advice and thought that rather than begging, I'd pitch it in terms of a topic you might post (or just comment) on. (sneaky, right?)

The issue is whether a hypothesis that predicts a certain region will be activated in two dissimilar tasks is appropriate for fMRI, given the pros & cons of conjunction analyses as debated by Friston & Price, Caplan & Moo, and Nichols et al.

Some would say that such a hypothesis is actually predicting a null effect - the absence of a difference between conditions. To my mind that is an older version of conjunction (e.g., the Friston & Price '97 model of a two main tasks-effects on a voxel and a lack of interaction). The newer Friston & Price 1999 "MS/GN" model, and the Nichols et al "MS/CN" model, retain only those voxels which show a significant effect in both tasks (allowing interaction effects to occur) and so it's not necessary to "accept the null."

Nonetheless, it seems that there might be other more powerful ways of doing conjunction analyses. For example, one might predict that the individual voxels V which compose clusters appearing in a conjunction map of task A & B are each correlated across tasks. That is, the activity of a particular voxel in task A can be predicted by the activity of voxels immediately surrounding V[i] in task B.

Any thoughts? Your previous posts discuss conjunction analyses without mentioning these problems, so I assume you're not persuaded by the argument that interpreting conjunction analyses as revealing common neural processes actually reflects accepting the null hypothesis. Obviously, the null hypothesis can only be rejected or not rejected.

At August 31, 2008 10:36 PM, Blogger The Neurocritic said...

Hi Chris. I'm not sure I'll have time for an entire post on conjunction analysis, since that would require a careful review of the [somewhat dense] literature. But I'll try to comment a little further in this space, in a little while.

At September 01, 2008 12:32 PM, Anonymous Anonymous said...

Thanks - I don't expect a treatise, I was just hoping you could say whether it's weak to propose an fMRI design that primarily predicts a conjunction effect in a particular region.

At September 01, 2008 5:00 PM, Blogger The Neurocritic said...

It might depend on what you want to test exactly, and if you want to follow Nichols et al. 2005 or Friston et al. 2005 (or some other method, as you said).

In other words, I'm not an authority on the matter.

From Friston et al. 2005:

Changes to SPM

The key contribution of Brett et al. and Nichols et al. is to enforce a re-evaluation of inferences required in the context of cognitive conjunctions. This has clarified the constraints on conjunction designs, namely that the components that are not common should be unique to a small number of contrasts, ideally u = 1. Furthermore, valid inference with the minimum statistic requires one to infer that k > u. This can be effected simply using the null distribution of the minimum statistic under the global null for n–u contrasts. A special case of incongruent contrasts, used in cognitive conjunctions, is when all the treatments are the same and the contrasts are congruent. In this instance, u = 0.

[Not that this quote comes as news to you.]

At September 01, 2008 6:46 PM, Blogger The Neurocritic said...

But overall I would say no, I don't think your proposal is weak.

At September 02, 2008 7:16 AM, Anonymous Anonymous said...

Thanks for your opinion neurocritic! Now I can respond to my nay-saying faculty advisors "but but but an anonymous blogger who really knows his shit says it's OK"


At September 02, 2008 7:41 AM, Blogger The Neurocritic said...

You're welcome. I'm sure that'll carry a lot of weight...


Post a Comment

<< Home

eXTReMe Tracker