GameBanshee Forums

http://news.bbc.co.uk/2/hi/health/4426184.stm

Now there are two reasons i have posted this. First to get CE's reactions and comments to the news article and research.

Secondly am i the only one who thinks this causal linkage is absurd. Loneliness: http://dictionary.reference.com/search?q=loneliness

The fact that you are alone, means people do not interact with you. The basic definition as per the dictionayr. That is a decision on the part of other human beings and thus can not based on your genome.

Many things can be influenced by genes even if they are not controlled by genes. It's two very different things. For example, if a certain gene makes you more likely to take contact with strangers than at group level people who have this gene will probably be less lonely than people who lack this gene. It's nothing strange about that.

I don't know anything about this particular study though.

I have the full article in front of me. There is nothing absurd at all with this study, it's a twin study using the established model for assessing relative contribution of genetic and environmental influence on something.

What is annoying is that popular media, as usual, in their eager to simplify and/or make a better story, twist both the results and the conclusion so what they report is finally something else than the researchers said.

"Loneliness may run in the family", BBC claims that the researchers have said. In the article, the researchers say there is a 48% genetic influence on the reported feeling of loneliness according to a specific scale, and that genetic multivariate analysis is needed to find out whether it is loneliness in itself, or a related trait such as temperament traits, negative affect or appraisal style, that is heritable and also how environmental influence foster or suppress the expression of these genetic influences.

I hope everybody realises this is miles from saying "loneliness may run in the family". Yes it may, but that's not what the researchers said. It's a good article, although not very interesting since it doesn't contain anything novel. I'll explain more later about this type of research.

This whole thing strikes me as another nature/nurture thing, but it also makes me think of the people who actually argue that only gay people have gay children. There may be some validity to the idea that genetics is partly-responsible for someone being lonely, but...

My parents, when they were younger, were quite sociable. They both had gaggles of friends, liked to be at parties, had fun in groups, etc. Some time during while I was a little kid, that stopped being so good to them. My mom started to despise (that's too strong a word but I can't think of anything more benign) one of her friends, who was a social butterfly. My dad, who was still quite sociable at work and work functions, wasn't known for being some kind of big party animal who was always socializing. In short, my parents became rather secluded and unsociable. We'll attend family functions, but it seems to me that it's more like it's a mandatory thing, not because they really want to go; not that they don't want to go, mind you. Anyway, my parents both stopped hanging out with friends and all their social activities were pretty much work-related.

As I grew up, I went from being quite sociable to being extremly unsociable. Was this the fault of genetics? After all, my parents weren't unsociable until after I was born; they enjoyed being sociable. Did I pick up their unsociability (is that even a word?) from them? Probably, especially since it developed later on in my life, about the same time they also were becoming unsociable. But then, on the other hand, I am quite friendly in person--despite my bitter and caustic attitude on- and offline--and I get along well with just about everyone (something I genetically picked up from my dad, no doubt ). So I think in my case, my instance of being "lonely" has more to do with the way I grew up than genetics. I tend to shy away from social gatherings and feel uncomfortable at them; I was afraid I wouldn't be comfortable doing retail work (I have since proven that I'm not only comfortable but I could produce a smile and friendly attitude even if my foot was being sawn off at that moment).

Methinks CE nailed it right on the head when quoting the "48%" deal. That's 50-50. So it can go either way, and in science, I'm pretty sure that's almost the same as saying nothing, or just saying they're unsure of the result. "Yes, we have a 50% chance of rain tomorrow, but because it rained yesterday and today, I recommend you bring your umbrellas."

Ok, this post is for those who are seriously interested in medical science, its’ investigation methods, questions, problems and possible conclusions. The BBC report of the study Fas linked to, is yet another sad example of how distorted scientific activities become in popular media. If you are only interested in claiming “science is so bad, my mysticistic personal feelings are much more reliable to gain knowledge”, skip this post because it will be long and tedious to read. If you are interested in how serious research can turn into the familiar headlines of “scientists found the gene for nose hair” or whatever, read on and I hope to be able to explain how medical science in general and behavioural genetics in particular, work.

Problems in medical science

All medical science has a common aim – to find treatment and ultimately cures for disease. All medical science has a common problem – disease is seldom simple and univariate. A disease, any disease, is a manifestation of a large number of different factors that interact with each other. These factors can be genetic, biological, environmental, social, cultural. Even a simple cold is the result of a complex interaction between the characteristics of the virus, characteristics in the individual (such as immune system effiency) and environmental factors such as climate (cold? Warm? Humid? Toxic?) and sociocultural factors (is it accepted that you stay at home from work a couple of days and get better? Or are you forced to perform physical work that may worsen the infection? etc)

To various degrees, all medical science has to find ways to start understanding these complex interactive chains. It is very rare that a cure for a disease is found by chance, or by trial and error. In almost all cases, to find a specific cure, we need to know the etiology, the mechanisms that cause a disease, and to gain knowledge about the etiology, we must know the patophysiology of the disease, ie the measurable characteristics of the disease.

Medical science starts from empiric observation: we have before us a patient, a person suffering from specific symptoms. We study the symptoms and try to find out: What is typical for this disease and not others? What is specific for this disease, and what is shared features with other diseases? Can the symptoms be related to a common cause? Can that give a lead where to look, which organs, which mechanisms, for patophysiology and etiology? What is cause and what is effect, and what is primary and what is secondary? When you start tracing down the answers to these question, you must use a set of systematic strategies.

First, some terms that are central in medical science. I have linked to Wikipeda and to some review articles for those who want a more elaborate explanation:

1. Phenotype – the manifestation as it can be observed, if it’s an individual it’s the characteristics and constitution of that individual, if it’s a disease it is the collected characteristics of that disease.
2. Co-variation – by empiric observation, we know that people with a strong social network are more healthy than people who live in social isolation. We don’t know any causality, we only know the covary, they occur temporally at the same time, in the same individuals.
3. Markers – if something covaries very strongly with a disease, so strongy so it almost always appears together with the disease, we call it a marker for that disease. High blood pressure is a marker for cardiovascular disease. A marker may or may not be involved in the etiology of the disease, that we don’t know. The marker may cause the disease. The disease may cause the marker. Or they may both be caused by an underlying background factor. Sickle cell anemia (SCA) is a blood disease. People who have it do not get malaria as easily as other people. SCA does not cause malaria protection. Malaria protection does not cause SCA. Both are caused by an underlying factor, a point mutation is the HBB gene.
4. Endophenotypes – are measureable and quantifiable components of a disease or trait, that are directly related to genotype. It’s a high quality marker,
5. Genotype – the individuals genes.

Depending on technological development, some areas much more understood than others. Least understood of all human diseases, are the nervous system diseases. The brain is by far our most complex organ. If you have tuberculosis, you can take a blood sample and isolate the tubercele bacteria. If you get a myocardiac infarction, you will have very similar symptoms to other people who got a cardic infarction. In neuropsychiatric disorders there are no tissue samples you can take, and symptomatology is more heterogenous than in any other field. The heterogenous phenotype of many diseases, is one of the main reasons why we study markers and endophenotypes when we try to find out the etiology of a disease. So when you see this stupid headlines reporting “Scientists have found the gene for nose hair”, don’t judge the merit of the study until you know that nose hair is not an important marker or endophenotype for a common form of cancer or neurodegenerative disease! Popular media almost never report why this or that is studied. In reality, you have to give a very precise and hypothesis driven rationale to even be allowed to perform any study at all on humans beings (expect if you are in the arts, they don’t have ethical regulations like medical science).

Twin designs and behavioural genetics

The reason why we study twins, is because it is helpful to decide to what degree a disease or a trait is genetic and environmentally determined. If something is 90% environmental, resources are better spend of finding out what environmental factors needs to be improved rather than performing big genome scans in thousands of people. If something is genetic, as Huntington’s disease which is 100% genetic (autosomal dominant heredity, meaning that if you are a carrier of the disease gene, it’s 50% chance your offspring will inherit the disease) the best strategy to find a cure is probably genotherapy rather than sociocultural intervention.
Twins can be monozygotic (MZ, identical) and share 99.9% of their genes, and they can be dizygotic (DZ, non-identical, ie as genetically similar as normal siblings). By comparing intra-pair similarity between MZ and DZ twins, and using shared environment and shared genes as factors, we can estimate the relative importance of genetic and environmental factors. For a review of how and why we can gain unique knowledge by twin studies and how these estimation models work, see this excellent review written almost in layman language:
http://www3.interscience.wiley.com/cgi- ... 4/PDFSTART
EDIT: please click "full text: pdf" above the title in order to read the full article.

Operationalised definition

In science, you can’t just take a term and claim to study that. You must specifiy exactly how you define what you study, so others can replicate your work. This is called operationalised definition (op def), and must be present in all scientific articles. You can’t just say you studies 50 patients with schizophrenia. Schizophrenia according to whom? According to clinical diagnosis by observation? According to the DSC-IV criteria? The ICD-10 criteria? The old Kraeplin criteria? Or your own home-make fantasy criteria? Thus, it is a cause of major misunderstand that popular media often fail to describe the op def of the studies they report. In behavioural sciences, where global phenotypes such as “intelligence”, “fear” or “loneliness” often overlap with normal language, it’s even more misleadning not to clarify the op def. Fas made no mistake when he looked up the definition of “loneliness” in a dictionary. BBC made a great mistake when wrote the header “Loneliness could be in your genes -
Loneliness may run in the family, researchers have suggested”. Later in the text, they refer to “experience of loneliness”, but they don’t clarify that they do not mean loneliness as in actually being physically or emotionally alone or isolated. In reality, Boosma et al (the authors of the Dutch twin study) defined “loneliness” as result on two questions in a self-report scale called YASR, that measures anxiety and depression.

to be continued in the next post

Summary of the article BBC reported

I cannot post the full article without violating copyright laws since it’s only available on a subscription site, so you have to trust my summary, go to a university library, or buy the latest number of Behavior genetics. If anyone is very interested in this fairly uninteresting study, e-mail me at C_Elegans@hotmail.com and I will send you the pdf. However:

The authors start by describing the state of perceived loneliness as a complex set of feelings of unmet social and emotional needs. The refer to several studies that have shown that perceived loneliness is related to self-esteem, mood, anxiety, shyness, social skills, social support, depression, anger and many other socio-emotional states. The authors went on and constructed an index for perceived loneliness by using a self-report scale called Young Adult Self Report (YASR). YASR is designed to evaluate emotional and behavioural problems, drug use and adaptive functioning, and includes eight areas: Anxious/Depressed, Withdrawn, Somatic Complaints, Thought Problems, Attention Problems, Intrusive Behaviour, Delinquent Behaviour, Aggressive Behaviour) and two broad-band scores (Internalising, consisting of Anxious/Depressed and Withdrawn ; and Externalising, consisting of Intrusive Behaviour, Delinquent Behaviour and Aggressive Behaviour). (Visser JH et al 2000, British Journal of Psychiatry,177: 59-65)

Boomsma et al made a two-stage factor analysis (a statistical method to group items) of the YASR and grouped together 2 items from the YASR “Anxious/Depressed”-scale. The two items were “I feel lonely” and “Nobody loves me”. Subjects could respond with one of the following alternatives: 0 = not applicable, 2 = a little or sometimes or 2 = clearly or often applicable. The individual score on this scale, Boosma et al called “loneliness”.

So, the 8 387 twins answered to these two items, and then Boosma et al compared how similar the MZ-pairs scored, and how similar the DZ-pairs scored. Then they used the established model to assess relative genetic and environmental influence on this score. You can read more about this method if you scroll down to the heading “Quantitative genetic analyses and heritability estimation” in this article. The conclusion was that 48% of the perceived loneliness was determined by heritable factors.

In the mandatory discussion part of an article, you must discuss your findings in a critical way, suggest interpretations and speculate in the meaning of your finding and suggest future research. Boosma et al suggest that the perceived loneliness as measured in their study, reflects a reponse pattern to outer stimuli, much like a personality trait (A personality trait is a general response pattern to stimuli that is stable over situation and time. Low-High Anxiety or Extraversion-Introversion are examples of personality traits that have shown very high stability cross-situation, cross-culturally and over time.) Like all studies personality traits, the perceived loneliness in this study, showed a heritability of about 50%. Thus, the authors discuss why such a response pattern would have persisted during evolution. (If it’s heritable, it must have since it still exists, that everybody understands?) They speculate that an individual who got isolated from the group for various reasons, would have had less chance to survive if it did not have a motivational drive to seek a group. The perceived loneliness would then act as an emotional “warning signal” (like anxiety, fear or other emotional signals) motivating the individual to seek company, and since individual who sought company had a larger chance to survive and reproduce than individuals who preferred to stay alone, the response pattern have persisted. It simply had a survival value.

Finally, Boosma et al discuss what their “loneliness” actually may mean, and report personality traits and behaviours that covary with perceived loneliness, such as anxiety, hostility, negative thinking, social avoidane, and also report that loneliness correlates highly to the personality trait Neuroticism (an anxiety trait). The conclude that further analysis is needed in order to investigate phenotype causation (ie Neuroticism causes loneliness or the other way around), and that genetic multivariate analysis is needed to find out whether it is loneliness in itself, or a related trait such as the above mentioned covarying traits, general negative affect or appraisal style, that is heritable and also how environmental influence foster or suppress the expression of these genetic influences.

My personal comments

Boosma et al’s perceived loneliness is composed of two items from a scale measuring Anxiety and depression. Anxiety and depressive personality traits are well studied and are known from several previous to have a 50% genetic influence. Furthermore, their loneliness correlated highly with the Neuroticism scale, one of the most validated and established anxiety scales that has been construced. Thus, I think Boosma et al are simply measuring a subscale of Anxiety, so their results are as expected.

Why study perceived loneliness? Well, it is known that perceived loneliness covaries with increased risk for suicide and a variety of psychiatric and somatic disorders as well as substance abuse. It is also well known that a strong social network is a protective factor for stress-related disorders, depression, and many other disorders. Thus, perceived loneliness may be a marker or even a predictor that can help identify groups or individuals that have a general increased risk for unhealth. Early identification means possibility for early intervention, and early intervention means better prognosis for almost all known diseases and disorders. However – this is frankly nothing new. There are hundreds, if not thousands of studies of anxiety and depression traits as a risk factor, including twin studies, so a study that shows that a subcomponent of anxiety and depression share the same features as the main component, is not the highest quality research one could imagine. This I believe is also reflected by the fact that the report was published in a relatively low-impact scientific journal. However, since I work at a front edge lab where we only perform studies that have a chance of resulting in gain of principally new knowledge, I am quite biased regarding the value of things that are already known

[QUOTE=Chimaera182]Methinks CE nailed it right on the head when quoting the "48%" deal. That's 50-50. So it can go either way, and in science, I'm pretty sure that's almost the same as saying nothing, or just saying they're unsure of the result. "Yes, we have a 50% chance of rain tomorrow, but because it rained yesterday and today, I recommend you bring your umbrellas."[/QUOTE]

No, 50% genetical influence and 50% environmental influence does not mean they are unsure of the result. It simply means that this particular trait/behaviour pattern/feature/disease is 50% heritable and 50% depending on environmental factors.

And @Fas: Bloody bastard, look what you forced me too you ruthless and cruel man, you owe me at least a great pasta when I visit NYC Or home made Pakistani

No, 50% genetical influence and 50% environmental influence does not mean they are unsure of the result. It simply means that this particular trait/behaviour pattern/feature/disease is 50% heritable and 50% depending on environmental factors.

Variables Galore, now at a store near you!

I cannot post the full article without violating copyright laws since it’s only available on a subscription site, so you have to trust my summary, go to a university library, or buy the latest number of Behavior genetics. If anyone is very interested in this fairly uninteresting study, e-mail me at C_Elegans@hotmail.com and I will send you the pdf. However:

CE, very few people here would be able to understand the terminology of an article that you are interpreting into laymen's speech onto here. However, your synopsis is useful and precise. I wish I could respond, but one of the reasons I spam is because it is the only thing I have time for inbetween timers going off and emails arriving and being sent.

What I was most interested in was buying behaviour genetics. I had no idea that was a possibility. Would make my job instantly easier.

Thanks CE. I have some comments which i will hopefully get to some time soon. I have read your comments and i have a question or two about the above study. Am I correct to assume that this study using MZ and DZ twins is based basically on a questionaire/survey? If that is the case did Boosma et al take into account the economic and social background of the twins? Lastly i can't access the blackwell synergy site. It won't give me access. Something about not having an ID. Any way around that?

I've tried to fix the link to Blackwell Synergy link now, please let me know if it still doesn't work.

Your questions:

1. Yes, Boosma et al:s twin study about "loneliness" was based on a 2-question self-report questionaire.

2. Socioeconomical factors are controlled for by comparing the twins intra-pair (ie between each other in the pair) and then compare the intra-pair measures between MZ and DZ twins. Since all twin pairs were reared together in the same family, it means intra-pair differences are not influenced by socioeconomical factors. Inter-pair (between-pairs, one pair compared to another pair) socioeconomical differences could be a factor if there was a systematic difference between the socioeconomic situation of MZ twins compared to DZ twins, but this has been demonstrated in many other studies that it's not the case. Whether parents gets MZ or DZ twins has no relationship to socioeconomic status.

Thanks for that explanation. I was asking if the study took into account the socio economic conditions of the different pairs in question. I see this study alot like any economic study. You assume everything else is constant - ceteris parabis - except for the factor you want to influence. The reason i ask if the socio-economic factors were taken into account is that whether a person has a healthy economic and social life style has a great influence on their notion loneliness.

Everything you say is true, but it doesn't relate to the study. The study uses x number of twins, who fill out a questionaire (which is subjective) and based on that extrapolates its conclusions. Now the conclusion that is "In the article, the researchers say there is a 48% genetic influence on the reported feeling of loneliness according to a specific scale" (taken from your first post). Now this could be considered valid if the economic and social factors were constant as well. You see where i am going with this? So did the study take into consideration the socio-economic background of each group of twins individually? I will comment more when i have the time.

Fas, I actually don't understand exactly where in this type of design you mean that socioeconomic factors could contribute to the result.

There were about 8000 twins in the study, 4000 MZ twins (2000 MZ pairs) and 4000 DZ twins (2000 DZ pairs).

You have 2000 MZ twins who fill out the questionairre. Then you calculate the difference between the individuals in the pairs. Let's say the mean difference between sibling 1 and sibling 2 in these MZ pairs are 10%. This measurement, ie the difference between a twin and his/her twin sister/brother, is called concordance. If 90% of the MZ twin siblings are equally lonely, and 10% of MZ twin siblings answer different from their twin sibling, we say loneliness has a 90% concordance, ie in 90% of the cases MZ twin siblings will share the same phenotype (in this case, the phenotype is loneliness).

Then you take the 2000 DZ pairs and do the same. Let's say the mean difference between sibling 1 and sibling 2 in the DZ pairs are 50%.

So, in our example, the concordance of loneliness is 90% in MZ twins and 50% in DZ twins. In other words, DZ twin siblings differ more from each other than MZ twin siblings do. Since both MZ and DZ twin pairs are reared together in the same family, sibling 1 and sibling 2 in each pair, has the same socioeconomic background.

So how could socioeconomic factors enter the equation? One way would be if MZ twins in general had a different socioeconomic status than DZ twins. But as I wrote above, this has been investigated, and there is no difference between the socioeconomic status of parents who get MZ twins or parents who get DZ twins. Another possibility would be if MZ twins later on in adult life, developed different socioeconomic behaviour than DZ twins. But this is not the case either. So I am sure how you mean socioeconomic factors would influence the estimate of 48% genetic effect.

PS. does the link work now?