They are definitely hypothesising that chemobrain is a somatoform disorder but they're using the term 'somatoform' in a way that I'm unfamiliar with; They say that the chemobrain patients are perceiving cognitive impairment, but because their perceived (subjective) cognitive impairment doesn't match what can be objectively measured in cognitive tests (i.e. cognitive tests are normal), it means that the symptoms are a false interpretation of reality (i.e. patients are perceiving impairment that doesn't exist), and so it is a somatoform disorder, despite other measurable biological abnormalities.
Technically the definition of somatoform does not automatically imply a psychological cause; it can be a physiological cause as well; though there is often an underlying assumption of a likely psychological cause, so to class a disease as somatoform tends to "tar" the disease as psychologically-caused.
But technically, somatoform is different to psychosomatic. Psychosomatic always means there is a psychological component to the symptoms (and always means the symptoms are objectively measurable); whereas somatoform can mean either a psychological or physiological cause to the symptoms (and always means the symptoms are
not objectively measurable; they are only subjectively detected).
But to conclude that the cognitive impairment isn't real seems like quite a jump in reasoning to me, as there could be a number of reasons why subjective experience doesn't match objective tests (e.g. the tests could be inadequate or they could be measuring the wrong thing. They acknowledge that cognitive studies have been small and underpowered.)
I have not looked at the paper in detail, but would first question the level of ME/CFS severity of the patients tested. Mild ME/CFS patients I imagine are not going to be too bad cognitively; we know this anyway, because lots of mild ME/CFS patients work. Whereas moderate and severe patients I expect are going to have far greater levels of cognitive impairment.
The other thing is nature of the cognitive tests. What are the tests measuring? I personally find that simple low-level cognitive tasks that are not complex (ie, do not have many components) I perform OK at. So if these were the type of cognitive tests used, that could explain why why subjective experience doesn't match objective tests.
But high level tasks that (a) require a bird's eye mental overview of the situation in order perform, and (b) have many complex interrelating facets, I perform very badly on.
Case in point is writing computer software. This was my profession, and I could write complex software blindfolded and with my hands tied behind my back. But now with ME/CFS, writing software is beyond me.
They have chosen - without explaining why, as far as I can see - to categorise mecfs as a non-pathological biological process, as opposed to a pathological one.
That choice itself would seem to be a subjective one. Just as some argue that for example ADHD or Asperger's are within the bounds of normal human experience. That is an arbitrary setting of what is normal and what is abnormal.
Seems a rather foolish statement, though, given that there is a huge spectrum of severity of ME/CFS, as well as ADHD and autism spectrum disorders.
One could possibly argue that very mild ME/CFS, where people still go to work, is within the normal bounds of human experience (I wouldn't argue this myself, but I can follow the argument). However, when you have a severe ME/CFS patient who is mostly bedbound, and extremely cognitively challenged, under no stretch of the imagination can you consider that normal.