I have been writing this week about mental illness denial, as a follow up to my debate with neurologist Fred Baughman on The Debate Hour. In my previous entries I pointed out that mental phenomena are manifestations of biological brain function, that sometimes classical diseases (like tumors and infection) of the brain can cause psychiatric symptoms, and further that brain function also depends upon more subtle biology – specifically the pattern of neuronal connections and the robustness of neurotransmitter activity, and that problems there can also result in mental disorders. Today I will discuss in more detail the evidence for a biological correlate to mental illness.

Dr. Baughman insisted numerous times during the debate that there is no evidence for a biological cause of any mental illness. This statement is demonstrably false – so let me demonstrate.

Of course, as mentioned above, there are numerous cases where classical pathology, the kind that Dr. Baughman admits into his privileged list of acceptable biological causes of disease, cause mental symptoms – such as depression, apathy, sedation, mania, and psychosis. Here there is no controversy – treat the underlying identifiable cause (if possible) and the mental symptoms should resolve, or at least improve or stabilize.

Epidemiological evidence

The controversy has to do with mental disorders for which there is no underlying identifiable classical pathology. Some of these disorders are identifiable syndromes, meaning that many patients present with a similar constellation of symptoms that take a similar course over time. For example, some people develop symptoms of visual or auditory hallucinations, combined with paranoid or delusional thoughts. They have an increasingly difficult time separating their own thoughts from external reality, may have ideas of reference meaning that they think everything somehow directly relates to them. For example they may think that the newscaster on television is speaking directly to them, giving them coded messages.

The above syndrome has long been recognized and is labeled as schizophrenia. Further, it has a characteristic epidemiology – patients tend to develop symptoms in their late teens and early 20’s, a bit later in females, they tend to wax and wane over their life but generally get worse over time.

There is good evidence that schizophrenia is strongly influenced by genes. Identical twins have about a 50% concordance, while fraternal twins have about a 4% concordance. This points strongly to genetic factors modified by environment, or requiring an environmental trigger.

There is a similar story with attention deficit hyperactivity disorder (ADHD). It is defined as a constellation of signs and symptoms, most notably a decreased capacity for sustained attention to a task, easy distractibility, and sometimes hyperactivity. Moreover, these symptoms must be greater than what is average/appropriate for age, and must exist in more than one setting – so if they exist only in school that is not sufficient.

There is also extensive evidence that ADHD has a strong genetic factor, from family studies, twin studies, and molecular genetic studies. Like schizophrenia, genes predispose to the disorder, about 50-60%, but environmental factors also play a role. For example, there is recent evidence that pre-natal smoking predisposes to ADHD. There is also evidence that links ADHD to an overall decrease in fetal development.

Therefore, both schizophrenia and ADHD are recognizable syndromes with a characteristic natural history, epidemiology, and evidence for genetic predisposition. This is sufficient to establish them as real entities, and is more evidence than we have for many medical conditions that do not involve mental symptoms.

The next question is what is the underlying cause or causes of these disorders.

Biological abnormalities

There is copious evidence for both schizophrenia and ADHD that brain function is abnormal. However, both of these disorders are actually clinical spectrums, not discrete entities. This means there are probably multiple underlying causes with similar clinical outcomes because the same brain function is compromised. (In other words, the clinical syndrome is determined by the brain function that is compromised, not the underlying cause – so multiple underlying causes are possible and even probable). Further, there are different identifiable subtypes of these disorders, and there are efforts to create more precise clinical descriptions that will allow researchers to reliably separate out the subtypes, and therefore more efficiently identify the underlying cause (because a subtype may have a single cause that will be easier to identify if it is isolated).

At this point the research points to multiple biological abnormalities in subsets of patients with schizophrenia and ADHD. Here is a sampling of the evidence.

In schizophrenia there is evidence that certain brain regions are smaller, on average, than in normal controls – for example the hippocampus and thalamus. There is also evidence from fMRI (images brain activity) studies that certain brain regions are hyperactive while others are hypoactive. Further many studies have shown that dopamine (a neurotransmitter) function is altered in the frontal lobes of schizophrenics.

To summarize, the exact types of abnormalities that we would expect to see are found in subjects with schizophrenia – namely a diminished volume of certain brain structures and abnormalities in the amount of activity in both connections and neurotransmitters. Moreover, these abnormalities are not found throughout the entire brain, or in random parts of the brain, but in those parts of the brain involved with the functions that are observed to be abnormal in schizophrenics, such as reality testing.

The same is true of ADHD. Parts of the brain involved in maintaining attention are underactivated in children with ADHD, such as the caudate as well as fronto-striatal and fronto-parietal circuits. These latter are involved in executive function and inhibition control. Other studies show that ADHD children have diminished reward anticipation.

Other studies looked at the volume of different brain areas and found that the relevant areas are smaller in ADHD vs controls.

Again we see that no matter how you choose to look at the brain – volume, fMRI activity, neurotransmitter activity – the areas of the brain that correlate to the symptoms of ADHD are small and/or dysfunctional.

Here is an excellent and up to date review of the current evidence for a biological cause of ADHD for those who are interested in further reading and references.

Of course (standard disclaimer to deniers) there is much that is still unknown about brain function in general and these disorders specifically. In any emerging and complex scientific field there will be gaps in current knowledge, sometimes significant gaps. Creationists, for example, will always have gaps in evolutionary theory to point to in order to maintain their denial.

How the deniers deny the evidence

Even with our current uncertainty and unknowns, the evidence I outlined above, which is just a taste of this vast body of research, taken as a whole is very compelling. The syndromes are reliably identifiable, they have biological and genetic markers, and they correlate to the types of brain dysfunction we would predict. There is a consilience of evidence here that is overwhelming. So what does Baughman, Szasz and other mental illness deniers say about this evidence?

One objection frequently raised is that, despite these studies, there is no biological diagnostic test for ADHD, schizophrenia or any similar mental illness. This, however, is an arbitrary, and even naïve, criteria to introduce. It may seem superficially reasonable – if there is hard evidence for a biological cause why isn’t there a hard biological diagnostic test? The reason is practical, not theoretical.

In order for a biological test to be useful for diagnostic purposes it must be cost effective and practical to administer, and it must have sufficient sensitivity and specificity – far above what is necessary to establish a connection scientifically. For example, in the studies I cited and most studies in schizophrenia and ADHD biological correlations are shown with around a 40-60% correlation. This is sufficient to establish a connection to a high degree of statistical significance. But it is almost worthless for clinical diagnosis, because applied to an individual it is almost a coin flip.

Further, such tests would not add significantly to the clinical diagnostic criteria. There is a good rule of thumb in medicine, before ordering a test you should decide how the outcome of that test will affect your clinical management. If I have a confident clinical diagnosis, a biological test that is only 50% sensitive, whether the result is positive or negative, is unlikely to affect my management.

In addition, the studies that are showing biological correlations, like fMRI, are mainly research tools and are not yet in widespread clinical use. So availability is limited, and resources are currently dedicated to research, not clinical studies. Clinical tests have to go through additional studies of validity and reliability in order to know how useful they will be clinically. This is additional information to the basic science studies that are designed to just establish a biological correlation.

Finally, it is worth noting that the same situation exists for many non-controversial biological diseases. For example, there are many studies showing early biological correlations of Alzheimer’s disease. But none of these tests are used clinically. Instead, the diagnosis is made by history, signs, and symptoms, and by ruling out other causes – exactly how ADHD and schizophrenia are diagnosed. So once again we see a false dichotomy being perpetrated by Baughman and others, when in reality mental illnesses are treated like many other non-mental or neurological entities in medicine.

Blame medication

Dr. Baughman and others have also refuted the copious evidence for a biological cause of mental illnesses by saying that the changes that are found (and they are acknowledging in this argument that the changes are real) are caused by the medications that are used to treat these disorders. So anti-psychotic medication causes all the changes we see in schizophrenia, and stimulants cause all the changes we see in ADHD. This claim is both implausible and now proven false.

First, it would be remarkable if a stimulant that affects the entire brain would cause anatomical and functional changes only in those parts of the brain that are involved with attention and executive function – those parts predicted by our current models of ADHD to be malfunctioning. Why don’t we see atrophy and hypofunctioning in other parts of the brain in these subjects?

But this is a theoretical argument – empiric data is more convincing. So to eliminate this possibility more recent biological studies have either focused on medication-naïve subjects (those never treated with medication) or analyzed them as a separate subgroup. So far studies of medication-naïve subjects (including some that I cited above) have all shown the same biological changes – disproving the medication hypothesis. This is not to say that medication has no affect on the brain, just that it is not responsible for the biological correlates that demonstrate these mental illnesses to have an underlying biological cause.

Tomorrow I will conclude this series of entries by discussing the practical aspects of diagnosis and treatment of ADHD and by answering questions that came up in the comments during the week.