Monday, December 19, 2011

Schizophrenia, Depression, the Hippocampus, and Information

19/12/2011 4:46PM

A review of a Nature Reviews Neuroscience paper, and then I just sortta wonder off and babble incessantly ... .

Nature publications generates some high quality reviews that are great for catching up on the latest developments in a given area. I've always  had a bit of interest in the hippocampus so was pleased that Nature Reviews Neuroscience offered the extensive review article below for free(each month the Review series and I believe some other Nature journals offer free articles).
Article:  A pathophysiological framework of hippocampal dysfunction in ageing and disease
Authors:   Scott A. Small, Scott A. Schobel, Richard B. Buxton, Menno P. Witter and Carol A. Barnes
Journal: Nature Reviews Neuroscience, Oct 2011.
Abstract | The hippocampal formation has been implicated in a growing number of disorders, from Alzheimer’s disease and cognitive ageing to schizophrenia and depression. How can the hippocampal formation, a complex circuit that spans the temporal lobes, be involved in a range of such phenotypically diverse and mechanistically distinct disorders? Recent neuroimaging findings indicate that these disorders differentially target distinct subregions of the hippocampal circuit. In addition, some disorders are associated with hippocampal hypometabolism, whereas others show evidence of hypermetabolism. Interpreted in the context of the functional and molecular organization of the hippocampal circuit, these observations give rise to a unified pathophysiological framework of hippocampal dysfunction.

The hippocampus and temporal lobes have always been regions of interest because many pathological conditions impact on these structures. In depression and Alzheimers these regions prove to be quite vulnerable but as the review points out there are important differences in the nature of neurodegeneration and metabolic function that occur in depression, Alzheimer's and normal ageing. This review was pleasing to read because it demonstrates how even over the last several years our understanding of neurodegeneration and CNS organisation continues to achieve ever greater degrees of resolution. Let us not be too optimistic though because as this review indicates at various points while fMRI is a very valuable tool the interpretation of BOLD results can be confounded by a range of issues that can be very difficult to factor into our analysis. The comments in the paper reminded me how simplistically we are presented with the results of fMRI studies and how easy it could be to create far too many "false positives". Yet another scathing critique of statistical analyses in imaging technology by the Greek mathematician John Ioannidis. I am no statistician but I have a deep fear that he is doing some very important work that may help to explain so many contradictory findings that appear in the literature. Here is a list of some links pertaining to his work.

Note the first publication in the list at the bottom of the page. An "epidemic of false claims". Reminds me of this wonderful study by the economist Dierdre McCloskey, with the cheeky title playing of the title of a famous paper written by the Nobel Laureate physicist Eugene Wigner in the 60's. "The Unreasonable Ineffectiveness of Fisherian “Tests” in Biology, and Especially in Medicine"(pdf download of full article)

Lies, Damned Lies, and Medical Science(Article from the The Atlantic)

His comments reminded me of what Jack Cohen did in the USA during the 60's. Do not so readily succumb to the "new phrenology" for while the clinical value of these imaging technologies is extremely important the analytic value of these imaging technologies needs to be approached with great caution. Their comments also remind me of some email correspondence long ago. A chap had mentioned on bionet.neuroscience that he had submitted a meta-analysis of various fMRI studies to journals which demonstrated the remarkable lack of consistency across the studies. His study was rejected by one and all. But they can't do that to Ioannidis, his reputation is too well established and he has a long tradition of demonstrating how parlous statistical analyses are much too often employed in biomedicine.

Now back to the Review but the above is important, and sadly it is the case that when reading all manner of studies the error rate as it currently stands involves far too much information pollution that has consequences not only at the analytic level and hence must be constantly borne in mind. Much more worrying are the clinical and public health implications. So hats off to John Ioannidis and perhaps now he could return to the home country and teach his fellow Greeks how to count.

The clear relationship between hippocampal dysfunction, depression, and Alzhiemers did not surprise me, the literature has always pointed in that direction, but I was very surprised to read the implication that dysfunctional hippocampal function may play a leading role in the development of schizophrenia and psychosis. The authors state in schizophrenia one study found that hypermetabolism occurs first in the hippocampus and is later evidenced in the orbitofrontal cortex, the implication being that the initial hippocampal hypermetabolism established deleterious metabolism in the orbitofrontal cortex and by implication a spreading wave of dysfunction across various regions of the CNS. The concept reminds of the famous statement by Hughling Jackson that the function of an cerebral region post brain injury may simply reflect a re-organisation of CNS function post injury. Here we see an example where in schizophrenia, often manifesting in a wide variety of behaviors and CNS alterations, by the time symptoms become the true initial causes may be long hidden from view.

It is going to be a long time before we understand schizophrenia. To be honest even as I think of the above findings my head spins with the incredible plethora of possibilities that can emerge when thinking about the various genetic, developmental, and environmental contingencies which can give rise to CNS pathophysiology. Not so long our genetic enthusiasm led to a number of headlines proclaiming a gene for this and a gene for that. We are now in the embarrassing situation of finding multiple genes typically being implicated in any given pathology but also being typically unsure as to whether or not said genes or genetic combinations and are essential preconditions for the appearance of the relevant phenomena. At present we have a truly overwhelming accumulation of data and paucity of understanding. The problem is immense but as this review amply demonstrates ongoing research is allowing us make more subtle and analytically relevant insights that will contribute not only to our understanding of the CNS generally but also to allow us to conceive of much more effective strategies to address various CNS pathophysiologies. For example, let's us assume it becomes well established that hypermetabolism of the hippocampus is a necessary pre-condition for the development of schizophrenia. In time it will become relatively easy to detect this and intervening at that point in time may prevent a sizeable number of individuals developing schizophrenia. Perhaps, my personal view on "schizophrenia" is much more cynical in that I think the diagnostic label is far too broad and this pollutes far too much neurobiologically based research into the condition. We are relying on diagnostic labels that were created at a time when our understanding of these conditions was akin to associating epileptic fits with demonic possession. Krapelin paved the way to a more empirically grounded approach to understanding schizophrenia, even labelling it as "dementia precox", which means premature dementia, a label consistent with his autopsy findings of schizophrenics which indicated noted and variable cerebral atrophy. though given the conditions under which the mentally ill lived at the time perhaps that cerebral atrophy was as much a product of the environment and general neglect of the poor bastards. In my view that label dementia precox should have stuck(but again the heterogenerity of the condition, studies on schizophrenics show widely variable CNS function and structure, and obviously shizophrenia has a wide behavioral repertoire). I suppose Hollywood could not have had much fun with such a technical sounding condition;  There are studies correlating schizophrenia risk with childhood deprivation, season of birth, and even migration. The beauty of the claims put forward in this review is that it is potentially testable. Thus when it is stated,

In the case of schizophrenia (FIG. 4a), one study found that hippocampal hypermetabolism occurs first, and that orbitofrontal hypermetabolism is observed only later in the disease course62. Furthermore, CA1 hypermetabolism was found to be associated with measures of delusions and hallucinations.


Such a study should be replicable, the findings should be very useful. The reference is to a study of 2009. The beauty of the finding is the important correlation that in schizophrenia there is a loss of GABA interneurons in the CA3 hippocampal region. GABA interneurons are inhibitory neurons and may be the primary reason for the hypermetabolism of the hippocampus even in the prodromal stages of the condition. However I find it difficult it to believe that such a discrete loss of GABA interneurons could be explained by some developmental, genetic, or environmental contingency set. All the moreso given the wide range of variance in schizophrenic symptoms and behaviors, not to mention general intelligence and risk factors. Sometimes I wonder if we gather so many "statistical ghosts" that we can all too easily lose sight of reality. Damn, I can even recall studies suggesting cells can change their neurotransmitter release from glutamate(excitatory) to GABA(inhibitory), suggesting a hitherto unanticipated neural plasticity that casts the discipline of neuroanatomy in a whole new light. Can we even meaningfully talk about a "static" neuroanatomy, must any neuroanatomical model incorporate a type of temporal account of process dynamics"? For example, it is not enough to know which cerebral regions are metabolically active and when but how the various causal cascades spread throughout the CNS. I'm not arguing the static neuroanatomical models are wrong so much as too limited in conveying an adequate impression of what happens "in the black box".

Even the synchronisation phenomena, as fascinating as it is, tell us nothing about what happens in the black box. In fact one would expect more a sequence of firing patterns, as if information is "flowing through" the "network", being analyzed at region X, which generates motor output Y, then all those neurons shut up and a new set come along for the next task. But nooo .... . I am so cynical of attempts to get inside the black box that when I think of firing neurons I only think of "responses". It is neutral, it doesn't suggest "number crunching" or "information processing" or "symbolic manipulation" or "maps of maps", it is incontrovertibly true that it is a response. I'm now ignoring the black box, I cannot think about number crunching or information processing or maps, I can only think about responses to contingencies leading to appropriate motor outcomes.

In my view, neurons are not "sending information". Neural synchronisation may present a hint of the precious neural code, that ghost don't know when to give up, but again it isn't telling us anything meaningful about information processing or signal transduction, although a friend of mine, Frank F Lefever,, likes to ponder the mystery of how neural synchronisation can be maintained with differing axonal lengths for signal transduction. That story was communicated to him by the Llinas, of thalamic 40 Hz chap. Striking how frequencies emerging from cortical activity appear so fundamental to CNS function, and how the presence of some frequencies can portend unfortunate days ahead. This again is a a synchronisation phenomenon. A study released this year may touch on why that is so. Unfortunately I've been slack this year and didn't archive it but the important past is that it overturns our conventional understanding and neuronal depolarisation and signalling(who, me, iconoclastic?). What they found, and I think this refers to GABA interneurons not glutamate neurons, was that a single GABA cell could receive many signals from other neurons but never generate an output. Then, seemingly spontaneously, it would start generating output for a full 30 minutes without input. Reminds me of Microsoft. So synchronisation could be maintained for extended periods of time through this GABA regulatory function? Perhaps, it was especially interesting to note that this year a study claimed, and I am so amazed at this I'm not accepting it just yet, that a single giant GABA interneuron was regulating the function of some 50,000 sensory neurons. Synchronisation may be achieved through a passive selection process(oh Mr. Edelman and his "neural darwinism" in reverse?), where GABA interneurons "shut down" those neurons and\or neuronal assemblies which are not the dominating synchrony in play or a synchrony of neurons that is not generating appropriate feedback.

So I am now back to depression, which suggests  my GABA interneurons aren't doing too well. But my working memory is fine. In the Review I was pleased to read that ongoing studies had validated a study on depression I long ago read but again had not archived so was relying on my diminishing memory ability. The study found that in depression people had hypermetabolic hippocampal function and that those who responded to the antidepressant drugs showed a reduction in hippocampal function. That item pleased me because at the time I was exploring a neuroimmunological analysis of depression and it became clear to me that the high density of glutamate receptors in the CA1 region of the hippocampus together with the hypermetabolism of the hippocampus would, over time, increase the general rate of cell death through excitotoxicity mediated by excessive glutaminergic function. Now I'm not entirely satisfied with that and the full picture includes a consideration of MR\GR receptors(steriods), cytokine expression profiles, BDNF prodn and histone acetylation implications of that, rate of neurogenesis in the dentate gyrus(complicated feedback with hippocampus there), inflammatory state. One hallmark of depression is temporal lobe atrophy, including the hippocampus. What really surprised me was the recovery, suggesting that under the appropriate physiological states neurogenesis can do a remarkably good job of facilitating recovery of hippocampal and temporal lobe health.

It seems counter-intuitive, the very word, "depression", suggests loss of function, but here we find that in recovery of major depression apparent loss of function in the hippocampus is a hallmark of recovery. Use it or lose it? In this case the exact opposite. There is a specific frequency associated with healthy hippocampal function, the number escapes me, might be theta waves. If you have lots neurons firing away, and consider the loss of CA3 GABA interneurons in schizophrenia, you're much less likely to have synchrony. There may be an upper limit to how many neurons can become sychronisations for tasks a,b, or c; all the moreso given that imaging studies suggesting widespread activation across the CNS during various tasks, axonal length differences alone should make synchronisation problematic. But then we come to Llinas and the possibility of GABA interneurons playing a key regulatory role in sustaining neuronal synchrony. Can music induce synchronisation? Is that why it generates so much rhythmic movement vertically or horizontally or diagonally or ... .

So when these cerebral regions are lit up like a phrenologist's Christmas tree that does not necessarily mean that information is being processed or symbols are being manipulated. It could represent a state of readiness. The neurons are not receiving information, they are responding to stimuli. What is the essential difference between a sensory neuron responding to an environmental stimulus and a CNS neuron responding to a signal received from other  neuron? The response is the same: depolarisation blah blah blah but with obvious qualifications for given neuron types and states. Maybe I'm going for a reckless monism approach here but I'll take my chances. Any neuron is responding to stimuli. If giant GABA interneurons can actually regulate the function of thousands of sensory neurons we need to seriously consider the possibility that GABA function is not just about preventing epilepsy but plays a fundamental role in CNS function with respect to more complex behaviors and as I'll latter indicate even the most basic of movements. Hmmm perhaps this is trivially obvious but GABA neurons don't figure prominently in sensory function but in signal regulation. Synchronised neuronal firing of motor neurons is obviously essential for movement involving limbs across the axes of the body and GABA neurons can do a very good job of regulating that; I even recall a very old study that mysteriously vanished from my archives which argued that in the earliest of back boned creatures it is the GABA interneurons which play a fundamental role in creating the rhythmic movement so often required for crawling and swimming. Nothing to process here, a simple feedback set up keeps the rhythm going. No need for bicomponent theory or even algorithms. Safe, no need to think about "neural codes" and especially that "electrical signal" crap. Damn that used to annoy me because it was short and stupid leap from electricity to computers and I'm not going there.

The big philosophical bugbear here is defining information for the purposes of understanding the CNS is very difficult if not impossible. It is impossible for me which is why I look for other approaches. I think it is a huge red herring. But we then come to an ontological quandary with respect to resolving the definition of information. Basically it comes down to a level of analysis issue and I'm still not convinced it is a make or break issue but I fear it is. Basically, how are we going to define information when we have to use information to define information? Call me cynical but I have a feeling that isn't going to work. Now in our particular case the problem is hugely compounded because we are trying to understand information as it pertains to the very organ we typically associate as the information organ. We use mathematics to understand gravity, what are we going to use to understand the information organ? Life can be a real shit. It's Christmas and I need to play a game.


1 comment:

Zanaprin said...

Zoloft is an exceptionally popular drug for helping to control anxiety, panic attacks, stress, and depression.