Tuesday, August 9, 2011

Identifying emotions 1

My hands are cold and clammy.
My chest is tight-it feels like the breastbone is trying to collapse in on itself.
I feel like there is a dense object in my stomach weighing it down.
My arms and legs feel shaky, although I can stand and walk fine.
My skin is crawling, as though the shakiness in my arms and legs is the skin shaking.
I don't feel as present in my head as usual. It is difficult to quickly shift focus and walk without running into people.
There's a pulling sensation at my temples, which contributes to not feeling present. It's as though my awareness slipped out of my head through the temples and is floating there, just outside.
I'm ruminating. I keep running over what triggered this feeling, and thinking about what I should say in response. Directing my thoughts back to work is nearly impossible.
The trigger was reading someone's mentioning grandmother's experience hiding during the Holocaust, and how they're upset when people trivialize the word Nazi.

What feeling is this?

Note: Part of the issue I have, which is common with Borderline Personality disorder, is identifying feelings  and thus appropriate responses.

Monday, August 8, 2011

Thoughts during Gunderson et al., 2011

I was at the hospital that the study was in! A few years before the study though. One of the people I met there said that Gunderson was her doctor there. She said she was nervous at seeing such a big shot specialist.

I also saw the ads for the study- I think both the BPD arm and the depressive arm. I thought about signing up for it, but I didn't want to deal with my parents. I think BPD is associated with not-very-good family relations, so I wonder if the study was skewed by whether the patients with the diagnosis could convince family members to participate or were willing to tell them at all. The authors note that their sampling method and corrections assume that the participation by patients with BPD diagnosis was not affected by characteristics of the family. I feel that this is probably a reasonable assumption, but as I said, I do have some doubts about it.

Most BPD patients are women: I feel as though perhaps the way girls are brought up contributes to the triggering or development of underlying BPD susceptibility.

Under cognitive I don't think I have many of the things listed. Perhaps thought distortions, although I suppose if I'm having that, I could be having any of the others and not know about it. Only slightly related, I was reading a blog by a woman who suffered horrific child abuse (not central to the blog, incidental as context for a 9-11 story) and realized (again?) that other people have different experiences than I do and see things differently. I try to be empathetic, but every so often things like this make me realize how little I know about how other people feel and think.

Right now I'm feeling kinda ok. I took a long time to read that article so I had time to process it. It's a relief that there is (probably) an underlying partially genetic cause, because that means its not my fault.

However, realizing that 5% of the population has it makes me feel less special. Having a disorder makes me special, it means that it's okay if I do things not as well as everyone else. But if it's actually basically normal to have this, I have no excuse.

I think the above is some of the thoughts I'm supposed to try to avoid focusing on. I think I'm supposed to say, if I don't feel well, I don't feel well, and I have to take that, as well as other things into account when making my decisions.

Thats it for now

Sunday, August 7, 2011

Article read: Gunderson et al., 2011

Title: Family Study of Borderline Personality Disorder and Its Sectors of Psychopathology

This article on borderline personality disorder (BPD) had 2 purposes: First, to see if there was some kind of heritability for the disorder, and secondly to evaluate different models for the definition of the disorder.

Prior work had shown that there was heritability, but there were problems with the studies and the body of work on the subject is less well-developed than for some other disorders. The authors did find that there was heritability. If you have an immediate relative with BPD, there is a 3 to 4 times higher chance of also having BPD. The base level of BPD occurrence in this study was 4.9%, and other estimates range from 2% to 6%. Most (not all) diagnosed patients of BPD are women, so all the initial subjects (probands) in this study were women.

For the evaluation of different models, they were looking at 3 of them. One is where there is this one disorder, which they term "BPD-ness", and it manifests in 4 different areas sectors. The sectors are Affective, Interpersonal, Behavioral, and Cognitive and are consistent across models. The next model has the four different areas having common genetic and environmental factors causing them. The third model basically has little relation between the four areas, and is a non-theoretical mathematical baseline, a Cholesky model.

What are the different sectors?
Affective is dealing with feelings. For this disorder it's things like depression, anger, and unstable moods.

Interpersonal is dealing with relationships. Here the main things are intense relationships with lots of drama and fear of abandonment.

Behavioral is dealing with actions. Self-harm is the big one here, but other impulsive behaviour is also implicated.

Cognitive is dealing with thoughts. Weird thoughts, dissociation, paranoia are all possible symptoms in this sector.

Using statistics, including the AIC and BIC I described in a previous post, the authors found that both the theoretical models are superior fits to the data than the mathematical decomposition model. This is good. We do have a clue, at least a little bit. The AIC was only slightly able to distinguish the two models, but agreed with the BIC which gave a strong indication that the one disorder of "BPD-ness" underlies all the diagnosis sectors instead of them being independent results.

On to results: BPD has an underlying disorder of which the different sectors are manifestations. It has a heritable, probably genetic component of about 50%, but no specific genes have been identified. These results support those found in other studies of BPD, heritability, and model.

There are some caveats, of which I find most important is the following: Not all possible sectors associated with BPD were assessed. Including other sectors could change the model significantly.
Of course, any unknown bias in the sample selection could mess things up too.

Gunderson et al. Family Study of Borderline Personality Disorder and Its Sectors of Psychopathology. Arch Gen Psychiat (2011) vol. 68 (7) pp. 753-762

Monday, August 1, 2011


AIC and  BIC are two similar, but not identical statistical concepts. They stand for the Akaike and Bayesian Information Criterion, respectively, and are two methods of determining which of two or more competing models is probably the best to use.

These criteria do not tell you if any of the models is actually good at predicting reality, just which are better or worse.

Definitions in math:
AIC = 2k -2*logL
BIC = k*ln(N) - 2*logL
where k is the number of parameters in the model, logL is the maximized log-likelihood, and N is the sample size.
Again, they are similar and based on the maximum likelihood estimate, but are penalized for number of parameters in different ways. Since you should almost always have a sample size larger than 7 (ln(8) > 2), the penalty for number of parameters is greater using the BIC.
(Adding parameters to a model means making it easier to tweak the model to fit the data-it's a bit like cheating if you don't have a good reason to add the parameters. That's why these two criteria penalize you for having additional parameters.)

In the paper I am currently reading, both AIC and BIC are based on the Yuan-Bentler T2* statistic and a Chi-squared distribution. The T2* statistic is a test statistic, or a function that combines many aspects of the data (such as mean, standard deviation, or number of samples) into one number. A distribution is the set of possible values of the test statistic, and how likely each of them are. The Chi-squared distribution is a particular common distrubition of known form that is used with the assumption of the errors in the sample being independent and normally distributed about zero.