Medical Hypotheses 80 (2013) 791–794

Contents lists available at SciVerse ScienceDirect

Medical Hypotheses
journal homepage: www.elsevier.com/locate/mehy

A serotonin hypothesis of schizophrenia

Arnold E. Eggers ⇑
SUNY-Downstate Medical Center, 450 Clarkson Ave., Brooklyn, NY 11203, United States
Kings County Hospital, 451 Clarkson Ave., Brooklyn, NY 11203, United States

a r t i c l e i n f o a b s t r a c t

Article history: Chronic widespread stress-induced serotonergic overdrive in the cerebral cortex in schizophrenia, espe-
Received 28 January 2013 cially in the anterior cingulate cortex (ACC) and dorsolateral frontal lobe, is the basic cause of the disease.
Accepted 9 March 2013 The concept of excessive serotonergic stimulation is supported by NMR spectroscopy; peripheral deple-
tion of phospholipids, serotonergic 5-HT2A receptors being linked to phospholipase A2; positron emission
tomography data with serotonergic ligands; and the fact that blockade of serotonergic 5-HT2A receptors
by atypical neuroleptics slows down the course of the disease. Disruption of glutamate signalling by sero-
tonergic overdrive leads to neuronal hypometabolism and ultimately synaptic atrophy and grey matter
loss according to principles of brain plasticity. Normal dopamine input to an impaired ACC causes posi-
tive symptoms. Frontal lobe hibernation causes negative symptoms and cognitive impairment.
Ó 2013 Elsevier Ltd. All rights reserved.

Hypothesis arachidonic acid [2–4]. Nuclear magnetic resonance (NMR) spec-

Chronic widespread stress-induced serotonergic overdrive in
the cerebral cortex in schizophrenia, especially in the anterior cin-
gulate cortex (ACC) and dorsolateral frontal lobe (DLFL), is the basic
cause of the disease.
Introduction
It has previously been hypothesized that excessive serotonergic
drive coming from the dorsal raphe nucleus (DRN) in response to
stress can disrupt the function of cortical neurons in schizophrenia
[1]. This theory will not be repeated in detail here; the basic ideas
are that the DRN is upregulated over a long, potentially permanent,
period of time in response to stress and that aberrant intensified
serotonergic drive in the cerebral cortex, especially in the ACC
and DLFL is the upstream cause of the psychiatric features of the
disease. The 5-HT2A receptor, the main cortical serotonergic recep-
tor, is linked to phospholipase A2; this means that stimulation of
serotonin receptors leads to removal of phospholipids from the cell
membrane as part of a signalling cascade. It was hypothesized that
serotonergic signalling in the brain is so intense that it leads to to-
tal body depletion of phospholipids in a manner analogous to a
consumption coagulopathy or a hemolytic anemia. This in turn
causes the well-documented peripheral features of schizophrenia,
such as decreased phospholipid content of red blood cells and poor
inflammatory response in the niacin skin test, which depends on
⇑ Address: SUNY-Downstate Medical Center, Box 1213, Department of Neurology,
Brooklyn, NY 11203, United States. Tel.: +1 718 270 2051; fax: +1 718 270 3840.
E-mail address: eggersa@aol.com
troscopy data in schizophrenia show several things: there are in-
creased levels of phosphodiesters, which is compatible with
increased activity of phospholipase A2; there are decreased levels
of phosphomonoesters, which is compatible with decreased activ-
ity of protein kinase C, which is one of the second messenger sys-
tems linked to glutamate signalling; and there is decreased
utilization of ATP in the frontal lobe, which corresponds to the
‘‘hypofrontality’’, i.e. decreased glucose metabolism, seen on posi-
tron emission tomography (PET) scan [5,6]. Disrupted glutamate
signalling leads to decreased action potential generation and hypo-
metabolism of the affected neurons.
Two important features of schizophrenia which any theory
must explain are cortical atrophy and the effectiveness of dopa-
mine blockade in treating positive symptoms.
Cortical atrophy in schizophrenia
Cerebral cortical atrophy is a feature of schizophrenia which has
long been recognized, having been documented in autopsy, pneu-
moencephalographic and computerized tomographic studies. A
more recent magnetic resonance imaging (MRI) study using man-
ual parcellation showed that volume loss is specific to grey matter
and preferentially involves certain areas, such as the anterior cin-
gulate gyrus, posterior superior temporal gyrus, and inferior frontal
gyrus [7]. Prefrontal and temporal white matter is spared. Reduced
grey matter volume is seen in first episode schizophrenics un-
treated with neuroleptics, which suggests that the disease has a
subclinical prodrome [8]. Normal skull size implies normal volu-
metric brain development at least up to the age of six, when the
brain reaches 95% of adult size; this means that, while genetic risk

0306-9877/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mehy.2013.03.013
792 A.E. Eggers / Medical Hypotheses 80 (2013) 791–794
factors are presumably already in place, the active disease process
probably starts somewhere between age six and the time of diag-
nosis in adolescence or early adulthood [9]. Individuals who are
clinically normal but at high risk because of a strongly positive
family history have reduced amygdalahippocampal volume
[10,11]. A meta-analysis of longitudinal CT or MRI imaging studies
showed that grey matter loss continues in severely affected pa-
tients for up to twenty years or more [9,11]. Marked cortical loss
correlates with poor clinical outcome [12]. Grey matter loss is,
therefore, a marker of whatever the basic disease process is and
it continues over time in patients who do poorly.
The neuropathology of schizophrenia has been exhaustively re-
viewed by Harrison [13]. White matter sparing suggests that axons
are intact, which implies that neuron cell bodies are spared as well.
Pakkenberg reported that the total neuronal cell count in the fron-
tal lobe is normal [14]. This study depends on sampling of zones
and, as the author points out, it does not rule out a small percent-
age of neuronal cell loss. According to Harrison, the most likely
explanation for cortical atrophy is a decrease in both neuron cell
body size and synaptic volume, which are interrelated because cell
body size is proportional to the size of the axonal and dendritic
tree. Electron microscopy on autopsy material is problematic and
the best evidence for synaptic atrophy (or ‘‘reduced neuropil’’)
rests on reports of reduced synaptophysin, a protein found in pre-
synaptic terminals, in the hippocampus and DLFL [13]. Loss of neu-
ropil under the name of ‘‘synaptic pruning’’ is the process that
explains normal loss of grey matter at adolescence, which at that
age is accompanied by increased myelination and gain of white
matter.
Is synaptic volume loss, which is the histological basis for grey
matter loss, the root problem in schizophrenia? Is the synapse the
point of impact of genetic risk factors, synaptic failure then causing
the other features of the disease?
Synapses, which are composed of dendritic spines and axonal
boutons, are not static structures. New dendritic spines arise spon-
taneously and continuously and grow toward and form synapses
with axonal boutons. Trachteneberg et al. carried out direct visual
imaging of the somatosensory cortex in mice in the region which
receives sensory input from whiskers in the mystacial pad [15].
In vivo synaptic morphology of fluorescent neurons was directly
visualized and photographed for periods of up to a month. Many
dendritic spines survived for less than one day, 60% survived for
8 days, and 51% for 30 days. Reducing sensory input by trimming
a selected number of whiskers significantly increased the ratio of
transient to stable synapses, i.e. decreased the survivability of
new synapses, but did not change synaptic density over this period
of time. Most studies of synaptic plasticity have been carried out
on the somatosensory and visual cortices in such a way as to dem-
onstrate the expansion of sensory fields which follows selective
and partial ablation of sensory input; this may not be a good model
for brain deafferentation in general. Jones and Pons reported on a
long-term study of monkeys that had had a deafferented upper
limb for periods of twelve to twenty years [16]. They found that
the cuneate nucleus, which received virtually no input because of
near-total atrophy of the dorsal root ganglion, was markedly
shrunken. However, 85–95% of neurons were still alive up to
twenty years later [17]. Neurons in the parts of the thalamus which
presumably represented the missing limb were also ‘‘visibly shrun-
ken’’. They interpreted this as a kind of subtle trans-synaptic atro-
phy. There was no electron microscopic or synaptophysin data, but
the general picture is reminiscent of schizophrenia: most neurons
survived but they were shrunken because of an atrophied axoden-
dritic tree. The general principle seems to be that synapses and
ultimately even some neurons require electrical stimulation in or-
der to survive.
It is suggested here that this is the process which explains cor-
tical atrophy in schizophrenia. The neuropathology of schizophre-
nia does not demonstrate neuronal degeneration in the ordinary
sense of the term; rather, it shows the more subtle effects of
chronic deafferentation. The schizophrenic brain falls into a hiber-
nation mode in which normal intracortical communication dwin-
dles. Synapses are not used and, therefore, are not replaced in
the process of normal synaptic turnover. The ACC and DLFL, which
are two of the key areas affected in schizophrenia, both receive
strong serotonergic innervation and are both hypometabolic
[18,6]. They also have strong reciprocal interconnections [19].
Chronic reciprocal failure of intercommunication leads to mutual
synaptic volume loss. The problem becomes structuralized and
ossified. Negative symptoms and cognitive impairment ensue.
The role of dopamine
The ACC is the main target of the mesolimbic dopamine path-
way [20]. A hypothetical explanation for hallucinations and other
positive symptoms in schizophrenia is that a normal dopaminergic
barrage hits an impaired target, which then malfunctions. This is
different from the standard dopamine hypothesis, which states
that an abnormally large dopaminergic barrage hits a normal tar-
get. More specifically, the revised dopamine hypothesis posits that
too much dopaminergic stimulation of a normal ACC (mesolimbic
pathway) causes positive symptoms and too little dopamine stim-
ulation of a normal DLFL (mesocortical pathway) causes negative
symptoms/cognitive impairment [21]. Increased dopamine release
would be expected to decrease the number of unoccupied dopa-
mine receptors available for binding of radioactive ligands in PET
studies and decreased dopamine release to do the opposite.
PET studies are technically demanding and, in the case of D2
receptors, not possible until recently for the cerebral cortex be-
cause of low receptor density. Studies are also limited by the diffi-
culty of obtaining adequate numbers of first episode
schizophrenics to do statistical analysis. Many of the D2 receptor
studies focus on the striatum, although it is not clear what the stri-
atum has to do with schizophrenia—the nigrostriatal system re-
lates to Parkinson’s disease and the mesolimbic and mesocortical
systems to schizophrenia. The striatum probably does not partici-
pate in consciousness and is probably not the source place of hal-
lucinations and delusions. Therefore, the focus here will be on
studies relating to pathways which are part of the dopamine
hypothesis.
With regard to the mesolimbic pathway, two groups have re-
ported decreased D2 receptor binding potential in the ACC, one
using an [18F] fallypride ligand and one an [11C]FLB 457 ligand
[22,23]. As noted in the discussion sections of these papers, the
ACC is an area of the cortex which is already shrunken in first epi-
sode schizophrenics. While it is possible that decreased receptor
binding potential is due to increased dopamine release, it is also
possible that it is a non-specific effect of atrophied synapses. The
situation is perfectly ambiguous.
With regard to the mesocortical pathway, a recent study of first
episode schizophrenics showed no difference in D1 receptor bind-
ing potential anywhere in the basal ganglia or neocortex in schizo-
phrenics compared to controls [24]. Another D1 study, using a
patient population consisting largely of chronic schizophrenics
who had been exposed to neuroleptics but had been drug-free
for at least fourteen days, did show increased D1 receptor binding
potential in the DLFL [25]. Interpretation of these studies is again
complicated by the fact that the DLFL has already undergone con-
siderable atrophy at the time of diagnosis.
In summary, the dopamine hypothesis has not been supported
by PET studies in the way that was originally anticipated. The fact
 A.E. Eggers / Medical Hypotheses 80 (2013) 791–794
that the key areas, the ACC and DLFL, are atrophic makes interpre-
tation of any results ambiguous. The classical dopamine hypothesis
assumes that the abnormal dopamine barrage hits a normal target.
The glutamate hypothesis, which posits an upstream cause for
dopaminergic abnormalities, suffers the same problem because it
makes the same PET scan predictions. Other problems with these
theories are that they fail to account for the following aspects of
the disease: (1) cortical atrophy, (2) relationship to stress, (3)
peripheral phospholipid abnormalities, and (4) the ability of sero-
tonergic blockade to slow down the course of the disease (see be-
low). It is, of course, true that neuroleptics treat positive symptoms
via D2 blockade in the ACC. The reason why first episode schizo-
phrenics already show grey matter loss may be that the disease
does not become symptomatic until the ACC is damaged.
Other data supporting a serotonin hypothesis
PET imaging has also been done with serotonergic ligands in
schizophrenia. Ngan et al. reported that, compared to age-matched
controls, first episode schizophrenics had decreased binding poten-
tial of [18F]setoperone, a 5-HT2A ligand, in the DLFL and ACC [26].
The authors believed their study was superior to two negative pre-
vious studies because they used voxel-based image analysis in-
stead of a region of interest approach. Decreased ligand binding
is consistent with increased occupation of receptors due to in-
creased serotonergic drive.
Conventional and atypical neuroleptics have different effects in
schizophrenia. In a two-year randomized double-blinded study of
first-episode schziophrenics treated with either haloperidol or
olanzapine, olanzapine was found to protect against progressive
grey matter loss for the first year while haloperidol did not [27].
Whole brain grey matter loss was 3.41% for the haloperidol group
and 1.72% for the olanzapine group (P < 001). Volume differences
correlated with Positive and Negative Syndrome Scale scores and
neurocognitive functioning; the data for the second year was unin-
terpretable because of patient drop-out. This study is important
because it shows that atypical neuroleptics, which block both sero-
tonergic 5-HT2A receptors and dopaminergic D2 receptors, treat the
basic disease process and not just the symptoms. Conventional
neuroleptics, which block only D2 receptors, treat only the symp-
toms. It is suggested that this is because atypical neuroleptics block
serotonergic overdrive, which is the root problem.
Some additional observations
Therapeutic recommendations can be made on the basis of this
theory. One of these relates to valproate, whose mechanism of ac-
tion has been debated for a long time. In contrast to negative re-
ports from other groups, Broicher et al. in a recent study of
selected drugs that block the T-type Ca2+ current (IT) in acutely-iso-
lated thalamic neurons in rats, showed that valproate could block
IT by 20–53% depending on the strain of rat [28–30]. This is impor-
tant because IT is the pacemaker current which drives the dorsal
raphe nucleus (DRN), which was the first pacemaker nucleus iden-
tified in the central nervous system [31,32]. Pacemaking is an
intrinsic feature of the DRN, apparent in isolated neurons, and
not a network property. In the previous paper, it was hypothesized
that serotonergic overdrive derives from excessive pacemaker
activity in the DRN in response to stress [1]. Therefore drugs such
as valproate, ethosuximide, topirimate, and zonisamide, all of
which are known to block IT, are logical add-on drugs for the treat-
ment of schizophrenia. Valproate also blocks the persistent sodium
current (INaP), which is a pacemaker current in the entorhinal cor-
tex [33,34]. It was previously hypothesized that there is a deleteri-
ous effect of the DRN on the hippocampus which is mediated via
793
stimulation of pacemakers in the entorhinal cortex [35]. There
are, therefore, two speculative sites of action for valproate which
are in tandem.
In summary, the serotonin hypothesis of schizophrenia explains
the origin of positive and negative symptoms, the relationship or
schizophrenia to stress, cortical atrophy, peripheral depletion of
phospholipids, and the effectiveness of dopamine blockade in
treating positive symptoms. It is supported by NMR spectroscopy,
PET data with serotonergic ligands, and clinical data showing that
serotonin blockade slows the course of the disease.
Conflict of interest statement
None.
References
[1] Eggers AE. Extending David Horrobin’s membrane phospholipid theory of
schizophrenia: overactivity of cytosolic phospholipase A2 in the brain is caused
by overdrive of coupled serotonergic 5HT2A/2C receptors in response to stress.
Med Hypotheses 2012;79:740–3.
[2] Horrobin DF. The membrane phospholipid hypothesis as a biochemical basis
for the neurodevelopmental concept of schizophrenia. Schizophr Res
1998;30:193–208.
[3] Berger GE, Smeny S, Amminger P. Bioactive lipids in schizophrenia. Int Rev
Schizophr 2006;18:85–98.
[4] Puri BK, Hirsch SR, Easton T, Richardson AJ. A volumetric biochemical niacin
flush-based index that noninvasively detects fatty acid deficiency in
schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 2002;26:49–52.
[5] Pettegrew JW, Keshevan MS, Panchalingam K, et al. Alterations in brain high-
energy phosphate and membrane phospholipid metabolism in first-episode
drug-naïve schizophrenis. Arch Gen Psychiatry 1991;48:563–8.
[6] Buchsbaum MS, Ingvar DH, Kessler R, Waters RN, Capplletti J, et al. Cerebral
glucography with positron tomography. Use in normal subjects and in patients
with schizophrenia. Arch Gen Psychiatry 1982;39:251–9.
[7] Yamasue H, Iwanami A, Hirayasu Y, et al. Localized volume reduction in
prefrontal, temporolimbic, and paralimbic regions in schizophrenia: an MRI
parcellation study. Psychiatric Res Neuroimaging 2004;131:195–207.
[8] Gur RE, Turetsky BI, Bilker WB, Gur RC. Reduced grey matter volume in
schizophrenia. Arch Gen Psychiatry 1999;56:905–11.
[9] Pol HEH, Kahn RS. What happens after the first episode? A review of
progressive brain changes in chronically ill patients with schizophrenia.
Schizophr Bull 2008;34:354–66.
[10] Lawrie SM, Whalley H, Kestelman JN, Abukmeil SS, Byrne M, et al. Magnetic
resonance imaging of brain in people at high risk of developing schizophrenia.
Lancet 1999;353:30–3.
[11] Puri BK. Progressive structural brain changes in schizophrenia. Expert Rev
Neurother 2010;10:33–42.
[12] Staal WG, Pol HE, Schnack HG, van Haren NEM, Sefert N, Kahn RS. Structural
brain abnormalities in chronic schizophrenia at the extremes of the outcome
spectrum. Am J Psychiatry 2001;158:1140–2.
[13] Harrison PJ. The neuropathology of schizophrenia. A critical review of the data
and their interpretation. Brain 1999;122:593–624.
[14] Pakkenberg K. Total nerve number in neocortex in neocortex in chronic
schizophrenics and controls estimated using optical dissectors. Biol Psychiatry
1993;34:768–72.
[15] Trachtenburg JT, Chen BF, Knott GW, Feng G, Sanes JR, et al. Long-term in vivo
imaging of experience-dependent synaptic plasticity in adult cortex. Nature
2002;420:788–94.
[16] Jones EG, Pons TP. Thalamic and brainstem contributions to large-scale
plasticity of primate somatosensory cortex. Science 1998;282:1121–5.
[17] Woods TM, Popken GJ, Jones EG. Reduction in volume of the cuneate nucleus
in adult monkeys following long-term deafferentation of the upper limb. Soc
Neurosci Abstr 1998;24:1512.
[18] Berger B, Trottier S, Verney C, Gaspar P, Alvarez C. Regional and laminar
distribution of the dopamine and serotonin innervation in the macaque
cerebral cortex: a radioautographic study. J Comp Neurol 1988;273:99–119.
[19] Paus T. Primate anterior cingulate cortex: where motor control, drive and
cognition interface. Nature Rev Neurosci 2001;2:417–24.
[20] Gaspar P, Berger B, Febvret A, Vigny A, Henry JP. Catecholamine innervation of
the human cerebral cortex as revealed by comparative immunohistochemistry
of tyrosine hydroxylase and dopamine-b-hydroxylase. J Comp Neurol
1989;279:249–71.
[21] Schwartz TL, Sachdeva S, Stahl SM. Glutamate neurocircuitry: theoretical
underpinnings in schizophrenia. Front Pharmacol 2012;3:195–205.
[22] Suhara T, Okubo Y, Yasuno F, Sudo Y, Inoue M, et al. Decreased dopamine D2
receptor binding in the anterior cingulate cortex in schizophrenia. Arch Gen
Psychiatry 2002;59:25–30.
[23] Buchsbaum MS, Christian BT, Lehrer DS, Narayanan TK, Shi B, et al. D2/D3
dopamine receptor binding with [F-18]fallypride in thalamus and cortex of
patients with schizophrenia. Schizophrenia Res 2006;85:232–44.
794 A.E. Eggers / Medical Hypotheses 80 (2013) 791–794
[24] Karlsson P, Farde L, Halldin C, Sedvall G. PET study of D1 dopamine receptor
binding in neuroleptic-naïve patients with schizophrenia. Am J Psychiatry
2002;159:761–7.
[25] Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, et al. Prefrontal
dopamine D1 receptors and working memory in schizophrenia. J Neurosci
2002;22:3708–19.
[26] Ngan ETC, Yatham LN, Ruth BTJ, Kuddke PF. Decreased serotonin 2A receptor
densities n neuroleptic-naïve patients with schizophrenia: a PET study using
[18F]setoperone. Am J Psychiatry 2000;157:1016–8.
[27] Lieberman JA, Tollefson GD, Charles C, Zipursky R, Sharma T, et al.
Antipsychotic drug effects on brain morphology in first-episode psychosis.
Arch Gen Psychiatry 2005;62:361–70.
[28] Broicher T, Seidenbecher T, Meuth P, Munsch T, Meuth SV, et al. T-current
effects of antiepileptic drugs and a Ca2+ channel antagonist on thalamic relay
and local circuit interneurons in a rat model of absence epilepsy.
Neuropharmacology 2007;53:431–46.
[29] Martella G, Costa C, Pisani A, Cupini LM, Bernardi G, Calabresi P. Antiepileptic
drugs on calcium currents recorded from cortical and PAG neurons:
therapeutic implications for migraine. Cephalalgia 2008;28:1315–26.
[30] Coulter DA, Huguenard JR, Prince DA. Characterization of ethosuximide
reduction of low-threshold calcium current in thalamic neurons. Ann Neurol
1989;25:582–93.
[31] Mosko SS, Jacobs BL. Recording of dorsal raphe unit activity in vitro. Neurosci
Lett 1976;2:195–2000.
[32] Burlhis TM, Aghajanian GK. Pacemaker potential of serotonergic dorsal raphe
neurons: contribution of a low-threshold Ca2+ conductance. Synapse
1987;1:582–8.
[33] Taverna S, Mantegazza M, Franceschetti S, Avanzini G. Valproate selectively
reduces the persistent fraction of Na+ current in neocortical neurons. Epilepsy
Res 1998;32:304–8.
[34] Alonso A, Llinas RR. Subthreshold Na+-dependent theta-like rhythmicity in
stellate cells of entorhinal cortex layer II. Nature 1989;342:175–7.
[35] Eggers AE. Redrawing Papez’ circuit: a theory about how acute stress becomes
chronic and causes disease. Med Hypotheses 2007;69:852–7.