Neuropathology and neurophysiology

The second trimester is a critical period for migration of neurones to the cortex and migration of dermal cells to form skin ridges. It seems that this migration does not occur normally in at least a percentage of cases of schizophrenia. This is illustrated by the finding of heterotopic clusters of nerve cells in the entorhinal cortex and the selective displacement of interstitial frontal white matter neurones. It has, however, been difficult to replicate such findings.

Tensor-based morphometry in a small study of childhood-onset schizophrenia over a five-year period revealed that this condition was associated with up to 2.2% slower annual growth rates in white matter, especially in the right hemisphere. 

Polysialic acid-rich neural adhesion molecule is involved in regeneration and plasticity of nerve cells and may be scanty the hippocampus in schizophrenia. Mice that are depleted in this molecule isoform 180 have abnormal neuronal migration and cytoarchitecture. Density of synapses peaks in childhood in the frontal cortex, decreasing by about one-third during adolescence due to progressive elimination of synaptic connections. Excess pruning has been suggested as a cause of hallucinations, as has abnormal co-activation of white matter tracts leading to confusion as to the source of inner speech. These suggestions are not without their detractors.

Schizophrenia might be a cortical or a subcortical disorder, the subcortex including the basal ganglia and thalamus where integration and co-ordination of neural function occurs.

A very recent study found fixed hippocampal volume loss in childhood-onset schizophrenia patients, findings not found in clinically unaffected siblings. A recent MRI study found corrected right hippocampal volume in first-episode, medication-naive schizophrenia subjects to be relatively small, and the lower the serum level of brain-derived neurotrophic factor (BDNF) the smaller was the hippocampus, possibly due to a toxic effect of psychosis onset on neurones. BDNF Met allele carriers have significantly greater decreases in frontal grey matter volume with reciprocal increases in the lateral ventricles and sulcal cerebrospinal fluid (CSF) than Val homozygous patients. According to McDonald the morphologies of schizophrenia and bipolar disorder are distinct but overlapping: schizophrenia is associated with subcortical and frontotemporal volume deficits, but grey matter volume is normal in psychotic bipolar disorder; and both disorders are characterised by volume deficits in white matter, with temporoparietal overlap. Lateral ventriculomegaly in schizophrenia seems to be more marked in males.

Results published in 1998 pointed to the non-progression of abnormal findings in schizophrenia (large cerebral ventricles, small medial temporal lobe structures, aberrant function of prefrontal cortex [PFC] and absence of gliosis) as evidence for a neurodevelopmental versus neurodegenerative process but not all studies agree, eg. certain workers suggest that some grey matter changes precede and others develop in association with first expression of psychosis. In an MRI study with an average of three years between scans, researchers found progressive reduction in frontal lobe white matter and increases in CSF volume during early-stage schizophrenia. Also, a 2006 study uncovered progressive brain volume changes on MRI, especially of grey matter, that were associated with clinical and functional outcome five years following a first episode of schizophrenia, and other, more recent studies support a neurodegenerative process. Nasrallah and Smeltzer (2002) favour a neurodegenerative process superimposed on abnormal neurodevelopment.

A very recent study found deterioration in cognitive functioning at baseline in clinically improved first-episode schizophrenia cases with further deterioration over a decade of treatment despite clinical improvement. The idea that any antipsychotic can be neuroprotective is still being debated. Murray believes that increased murine striatal D2 receptors may cause schizophrenia-like deficits in behaviour and cognition. A 2009 study looked at the effects of haloperidol versus atypical drugs on cognition in schizophreniform and first-episode schizophrenia patients and found that all of the drugs moderately and equally improved performance and such improvement was weakly related to changes in Positive and Negative Syndrome Scale (PANSS) scores.

Cerebral laterality may be decreased in schizophrenia, although there are discrepant findings. Crow suggested that psychosis is a side-effect of human evolution because of variation in genes controlling hemispheric asymmetry: because the corpus callosum cannot handle everything, language evolved by a process of hemispheric specialisation – phonological output sequence became localised to the dominant hemisphere, and its associations (signfields) were lateralised in part to the other hemisphere; thoughts are translated through a bi-hemispheric interaction into phonemes/speech by the speaker in frontal association areas, and decoded back into concepts/meanings by the listener in occipito-temporal-parietal areas. Crow believes the patient with psychosis cannot distinguish phonemic signals generated by the hearer from his own thoughts, or from signals that he receives from an interlocutor: his brain thinks that his own thoughts have an external origin.

Liddle suggested that human minds have become fairly independent of the external world, allowing for a more flexible response to a changing environment by employing stored memories to construct theories about the future – reality distortion derives from a confusion of ‘mental representations triggered by current external circumstances’ with ‘representations of past situations and representations of hypothetical situations’.

Electrophysiological studies employing perceptual tasks have demonstrated changes in auditory event-related potentials (ERPs) that differentiate patients with schizophrenia from controls. 

Prolonged P3 latency and reduced P3 amplitude indicate impairment of auditory processing in some people with schizophrenia which is independent of the presence of acute psychotic symptoms and uninfluenced by antipsychotic medication. Auditory P3 amplitude may decrease with illness duration. Despite frequent sparing of the P3 component of the visual ERP in such cases, there also is disturbance of visual attention in schizophrenia as shown by reduced frontal P2a and posterior N2b event-related components. Research with ERPs suggests a neurophysiological impairment in the maintenance of selective attention and the cognitive processes associated with target detection in schizophrenia, most likely caused by frontostriatal dysfunction. Work using visual target and competing visual stimuli has been interpreted as arising from anterior cingulate dysfunction. Also, functional magnetic resonance imaging (fMRI) during word generation by schizophrenia patients suggests an inability to inhibit competing cognitive processes. Researchers have reported reduced P3 amplitude in close relatives of schizophrenia cases and in people with schizotypal disorder. Other reports suggest that reduced amplitude of the P3 component on the left side only may characterise early schizophrenia.

Decreased neuregulin 1-induced activation of phosphoinositide 3-kinase/AKT system has been reported in association with impaired sensory gating in first-episode schizophrenia. 

EEG coherence measures have demonstrated a relationship between left frontotemporal connectivity and reality distortion, particularly in men.

A 2004 study used a conventional auditory odd-ball task in men and women with chronic or first-episode schizophrenia and matched controls: phase synchronous gamma activity was extracted from the EEG; chronic cases, particularly women, showed decreased global functional connectivity in the form of lower gamma phase synchrony. While first-episode cases showed a general decrease in the speed of frontal connectivity, the speed of global connectivity was faster in female patients. These findings suggest that females with schizophrenia experience further breakdown in cerebral network connectivity with chronicity of disease.

Other findings include reduced gamma activity and failure of lateralisation of activity to the right hemisphere during masking. A decrease in EEG-evoked responses in the gamma band when transcranial magnetic stimulation is applied to directly stimulate the frontal cortex suggests a possible intrinsic problem in frontal thalamocortical connectivity. Supporting thalamic dysfunction is all-night EEG recording, which shows that schizophrenia patients have a reduction in spindles during non-REM sleep that is unlikely to be due to medication.

A 2004 study found that clozapine improved P50 gating more than did other atypical or typical antipsychotics, but a controlled study of auditory ERPs in unmedicated schizophrenic outpatients found P50 gating to be normal. Dublin-based research found a deficit in early visual processing in well first-degree relatives of people with schizophrenia. It has been suggested that multiple visual information processing deficits derive from dysfunction of the magnocellular pathway, causing impaired attentional modulation of perceptual organisation and of natural image organisation.

Excitatory lateral connections in early stage visual cortical processing may contribute to complaints such as unclear seeing, partial or skewed sight, disrupted rectilinearity and abnormal figure-ground segregation. In chronic schizophrenia there is evidence of increased skin conductance activity at rest, and in socially demanding conditions the skin conductance level and variability was increased in the right hand. Asymmetric skin conductance may be a characteristic of the chronic from of the illness but skin conductance may also increase in the weeks preceding relapse or exacerbation of schizophrenia.

Eye-tracking abnormalities are not specific to schizophrenia, having been reported in a number of psychiatric disorders, although they appear to be inherited with schizophrenia. Over 80% of people with schizophrenia do have abnormal smooth pursuit tracking, with about one-third of their relatives having similar problems. 

Saccadic abnormalities are reported in schizophrenia. Both these and smooth pursuit abnormalities seem to be present at the start of the disorder. There may be an abnormal frontostriatal network that normally suppresses automatic eye movements and/or an abnormality of the frontal eye field. Saccadic abnormalities seem to be more common in people with schizophrenia than in their relatives, but more common in the latter than in people with mood disorders, who in turn are more affected than healthy controls. Saccadic task performance is more stable over time in schizophrenia than in bipolar cases, which supports a trait-like phenomenon in the schizophrenia group. 

A recent study of patients with schizophrenia and their siblings found elevated memory-guides saccade error rates relative to controls, reflecting failed inhibition of reflexive saccades to a visible target.

Concerning backward masking, patients with schizophrenia find it particularly difficult to identify the first/target stimulus. This may be due to failure to adequately activate the lateral occipital complex. Backward masking has been reported in some cases of chronic mania.

Neuroimaging

As a generalisation, decreased blood flow to the dorsolateral prefrontal cortex (DLPFC) correlates with negative symptoms while frontotemporal abnormalities correlate with positive symptoms including thought disorder. Schizophrenia subjects fail to activate the PFC cortex when looking at an emotive picture. Relative enlargement of the lateral and third ventricles and cortical sulci found on CT scans in some schizophrenia patients has been interpreted as being static and due to a neurodevelopmental rather than a neurodegenerative phenomenon. CT findings also include small brain volume and length, particularly on the left side and in males. Such findings seem to be independent of duration of psychosis or medication. Individuals at ultra-high risk of psychosis have been reported to have reduced grey matter in frontal regions bilaterally and, specifically, those who do become psychotic have less grey matter volume in the left parahippocampal cortex.

MRI work published in 2005 examined the hippocampi of young male schizophrenia cases, older chronic patients and matched control: schizophrenia patients had relatively small hippocampi, especially if chronic; younger patients had larger hippocampi if they had been treated with atypical drugs rather than haloperidol. A later study found, relative to healthy controls, bilaterally reduced hippocampal volumes in prodromal schizophrenia; only in those considered to be in the late prodrome did these reductions correlate with poorer performance on a test of delayed recall. Functional MRI has disclosed abnormal cerebral blood volume (CBV) increases in the CA1 subfield of the hippocampal formation and orbitofrontal cortex and abnormal CBV decrements in DLPFC; baseline CA1 CBV abnormalities predicted progression to psychosis from the prodrome; CBV in CA1 correlated with psychotic symptoms; and the results were not deemed secondary to medication.

McDonald et al (2004), using MRI, reported that genetic risk for schizophrenia was specifically associated with distributed grey matter volume deficits in bilateral frontostriatothalamic and left lateral temporal regions; genetic risk for bipolar affective disorder was specifically associated with grey matter deficits only in right anterior cingulate gyrus and ventral striatum; genetic risk for both disorders was associated with reduced volume of white matter in left frontal and temporoparietal regions; and these findings were viewed as indicating a common frontotemporal disconnectivity.

An examination of first-episode schizophrenia patients and matched controls with diffusion tensor imaging (DTI) and voxelwise analysis showed four clusters where fractional anisotropy values were significantly lower among patients; these were localised bilaterally to areas of white matter corresponding to superior and inferior longitudinal fasciculus, forceps major, anterior and superior thalamic radiation, and corpus callosum; and the authors concluded that decreased white matter integrity is present early and is localised in the fascicule that connects brain regions implicated in schizophrenia.

A 2009 study found that schizophrenia was associated with increased fractional anisotropy in corpus callosum, cerebral peduncle, left inferior fronto-occipital fasciculus, anterior thalamic radiation, right posterior corona radiata, middle cerebellar peduncle and right superior longitudinal fasciculus; increased fractional anisotropy was detectable in inferior sections of the corticopontine circuit; and the authors suggested that their results indicate extended cortical-subcortical changes in white matter integrity in schizophrenia and corroborate earlier work demonstrating white matter structural deficits in mainly long-ranging association fibres.

DTI and neurocognitive assessment in first-episode schizophrenia patients revealed that executive and motor deficits are associated with reduced white matter integrity in the main bundles connecting frontal and temporal cortices and in cortico-subcortical pathways. A recent meta-analysis of DTI work looking at the corpus callosum in schizophrenia suggests reduced fractional anisotropy in the splenium, indicating difficulties with posterior interhemispheric connectivity. Combined fMRI and DTI investigation of decision-making under uncertainty in schizophrenia found significant deactivation in a frontostriatocingulate network and increased radial diffusivity in temporal white matter that was negatively correlated with activation in parts of the frontostriatocingulate network. A meta-analysis of voxel-based morphometry studies found that the most consistent differences were relative deficits in the left superior temporal gyrus and the left medial temporal lobe.

Longitudinal MRI (baseline and 1.8 years later) was conducted with very high-risk people, first-episode psychosis cases and community recruits. Results suggest progressive grey matter reductions of the superior temporal gyrus during transition to psychosis. However, another study, while finding smaller superior temporal gyri on MRI in those at risk of psychosis, found no difference between those who did and those who did not go on to become psychotic.

Fractional anisotropy deficits in left temporal lobe and right corpus callosum may be more severe if there are no known relatives with schizophrenia.

Using the Wisconsin Card Sorting Test (WCST) to activate prefrontal areas, most but not all authors have found diminished blood flow in these areas while the task is being performed (‘hypofrontality’). The normal reaction to the WCST is an increased blood flow to the DLPFC. Hypofrontality can be missed if the patient is assessed while resting; it is best performed during an attention task. Some researchers find greater hypofrontality in chronic cases. Meyer-Lindenberg et al reported evidence in 2005 for abnormal hippocampal formation-DLPFC connectivity during working memory activation in schizophrenia.

A meta-analysis of the literature found that schizophrenia was associated with abnormal patterns of brain activation during both encoding and retrieval of memories. The PFC and hippocampus, among other regions, are implicated in the abnormal memory functions. Another meta-analysis found prefrontal activation deficits during episodic encoding and retrieval in patients with schizophrenia. A further meta-analysis of functional neuroimaging studies of executive function in schizophrenia found that healthy adults and patients with schizophrenia activate similar networks (DLPFC and anterior cingulate cortex [ACC]) but that patients show changed activity with deficits in DLPFC, ACC, and mediodorsal thalamic nucleus; increases in activity in other PFC areas may be attempts to compensate for deficits.

Recent MRI/fMRI study findings suggest that cognitive impairment is not related to brain structure but is a functional (hypoactivation during an n-back task) problem: cognitively impaired patients were structurally the same as cognitively intact cases.

A 2007 study discovered microstructural disruption of white matter in frontal, temporal and occipital white matter. A later study reported diffuse deficits in a number of types of white matter tracts and an inverse relationship of fractional anisotropy values with positive symptom scores in association fibres, interpreted as supporting a disconnection explanation for positive symptoms in schizophrenia. Widespread structural disconnectivity, including the subcortical region, was found in neuroleptic-naive, first-episode, Chinese schizophrenia patients using DTI. DTI and structural MRI in both schizophrenic subjects and controls demonstrated a relative reduction of fractional anisotropy in those parts of the corpus callosum that connect both frontal regions. A 2009 DTI study revealed the same white matter abnormalities in bipolar and schizophrenic patients compared to controls. Kanaan et al (2009), in a relatively large study, found widespread clusters of reduced fractional anisotropy affecting most major white matter tracts in patients and these were not associated with duration of illness or duration of treatment.

Recently, using an affective face recognition paradigm and fMRI, researchers found that patients with schizophrenia performed slowly. Comparison subjects recruited expected cortical areas more than did patients and more severely symptomatic patients showed relatively diminished recruitment. Increased symptoms correlated with excess amygdala and orbitofrontal cortex response to threatening faces, and comparison subjects showed a negative relationship between amygdala and cortical areas involved in cognition, while patients exhibited weakening of this relationship. In other words, the limbic system is over-responsive and the cortical facial recognition memory response is decreased.

Patients with auditory hallucinations were found to have reduced interhemispheric connectivity in both primary and secondary auditory cortices on fMRI compared to both non-hallucinating patients and healthy controls while passively listening to words. Does the patient ‘hear’ a voice coming from one hemisphere and interpret it as originating from outside of the self?

See part one of this series.

See part three of this series.

Declaration of interest: none