The Neural Correlates of Negative Symptoms in Schizophrenia: Examples From MRI Literature

Structural Brain Imaging Findings

Neuroimaging studies report inconsistent neural correlates of this heterogeneous construct. Negative symptoms have been associated with brain structure abnormalities in several frontotemporal regions, as well as in thalamus and cerebellum.^2,3 Associations of brain structural abnormalities with negative symptoms have also been found in antipsychotic naïve patients,^4,5 although not in all studies. ⁶ The majority of studies report associations of smaller brain volumes and more severity of negative symptoms; however, an opposite pattern of larger brain volumes being associated with increasing severity of negative symptoms has also been found. ⁷ Differences in study population, study design, and MRI methodology can account for some of the heterogeneity of results. An association between negative symptoms and thinner cortex in prefrontal regions was recently confirmed in meta-analysis within the ENIGMA consortium, which included 1985 patients with a schizophrenia spectrum diagnosis from 17 different study groups. ⁸ Only prefrontal cortex regions were investigated in this meta-analysis.

Functional Brain Imaging Findings

The hypothesis of dysfunction in the dopaminergic reward systems in schizophrenia has a long history where the ventral striatum has a central role where a meta-analysis ⁹ showed this region to be hypoactivated during reward anticipation and that this hypoactivation was related to negative symptoms. While patients with schizophrenia have a relatively intact capacity of feeling pleasure and consume rewards they display abnormalities in the anticipation of reward, that is, anticipating hedonic experience in the future. Using functional magnetic resonance imaging (fMRI) during a monetary incentive delay task it was found that unmedicated patients with schizophrenia displayed a striatal dysfunction. ¹⁰ The reduced activation in the left ventral striatum during anticipation of a probable reward was associated with negative symptoms. However, using a similar task Nielsen et al ¹¹ found a correlation of activity changes in the same regions to positive psychosis symptoms (ie, symptoms including delusions and hallucinations). Using a rewarded perceptual decision-making task, no direct correlations between activity in the ventral striatum and negative symptoms were obtained but altered functional connectivity between the ventral striatum and interior frontal gyrus was found to be associated with negative symptoms. ¹²

Gray Matter/Cortical Thickness

Results from studies investigating gray matter differences in patients with and without DS (DS vs NDS) or PNS (PNS vs non-PNS) have not been consistent. In studies examining gray matter volume differences in selected regions of interest between DS and NDS, findings include for instance smaller gray matter volume in regions such as the right temporal lobe, ¹⁷ the superior frontal gyrus and the middle temporal gyrus, ¹⁸ and the left nucleus accumbens. ¹⁹ Some of the heterogeneity of findings between studies may be due to small study samples and different regions of interest investigated. When investigating whole brain gray matter differences between DS and NDS patients using voxel based morphometry, Cascella et al ²⁰ identified several regions where DS patients had smaller gray matter volume compared with NDS, including the superior frontal gyrus, bilateral superior frontal gyrus and superior temporal gyrus, left supplementary motor area, left anterior cingulate, left insula, left cuneus and right putamen. In another voxel-based morphometry study, Lei et al ²¹ reported significant gray matter reductions in the left insula in drug naïve first episode schizophrenia patients with DS compared with NDS. Gray matter volume of insula was negatively correlated with severity of negative symptoms. Bodnar et al ²² investigated cortical thickness across the cortical mantle, and found that first-episode psychosis patients with persistent negative symptoms had thinner cortex in regions including the right superior temporal gyrus, right parahippocampal gyrus, and left orbital frontal gyrus compared with first-episode psychosis patients without persistent negative symptoms. In contrast with these findings, Voineskos et al ²³ found that both DS and NDS patients had thinner cortex in overlapping regions compared to healthy controls, and no significant differences between DS and NDS were found.

White Matter/Diffusion Tensor Imaging

White matter abnormalities in DS patients have been indicated by several diffusion tensor imaging (DTI) studies.^23-25 Lei et al ²⁶ performed both voxel based morphometry and DTI in first-episode antipsychotic naive schizophrenia patients diagnosed with either DS or NDS, their first-degree relatives and healthy controls. They found that DS patients had smaller extranuclear white matter volume compared with NDS patients, and white matter volumes in this region were negatively correlated with severity of negative symptoms. Interestingly, bilateral extranuclear white matter volumes were also smaller in relatives of DS patients compared to relatives of NDS patients. Whole brain comparisons of fractional anisotropy (FA) did not reveal any significant differences between DS and NDS, however, when mean FA was extracted from the uncinate fasciculus (a region where DS patients had lower FA compared with healthy controls), mean FA was significantly lower in the uncinated fasciculus in the DS compared to the NDS group. Abnormalities in the uncinate fasciculus have also been related to DS or negative symptoms in other studies.^23,24,27 Other regions where white matter abnormalities have been indicated in DS include the right superior longitudinal fasciculus ²⁵ and the right inferior longitudinal fasciculus.^23,25

Brain Neurodevelopmental Markers

DS and PNS are associated with poorer premorbid adjustment, an insidious onset, and longer duration of untreated psychosis, ¹⁶ which suggest abnormalities in early neurodevelopment. Takahashi et al ²⁸ investigated selected gross brain markers of neurodevelopmental pathology in in 38 patients with deficit schizophrenia, 39 patients with NDS and 59 healthy controls. Compared with NDS and healthy controls, DS patients had shorter adhesion interthalamica. Furthermore, alterations in OFC sulcogyral pattern were observed in DS, and not in NDS when both groups were compared with healthy controls.

Functional Magnetic Resonance Imaging

In a blood oxygen level–dependent (BOLD) study using a monetary incentive delay task comparing DS and NDS patients, brain activations showed reduced activity in the left dorsal caudate in the DS group. ²⁹ The dorsal caudate was recruited in the task among healthy controls in the anticipation of reward but not among the schizophrenia patients as a group. Further investigating this, the authors made a median split on avolition among the whole group of patients where it was showed that high-avolition patients showed reduced activity in the dorsal caudate compared with both healthy controls and low-avolition patients. All DS patients were in the high-avolition group but only 3 out of 17 of the NDS were in the same group. No activation differences between the groups in the ventral striatum were obtained.

Avolition-Apathy

From behavioral and neuroimaging studies emerges a complex picture of deficits in anticipation of pleasure, reward valuation and prediction, effort computation, and cognitive decision-making processes underlying motivational deficits in schizophrenia.^30,32-34 Results from studies specifically investigating the neural correlates of apathy in brain disorders support that apathy is associated with abnormalities in prefrontal-subcortical circuits involved in motivation and goal directed behavior. ³⁵ Apathy has been associated with smaller frontal lobe volumes, ³⁶ and more specifically, thinner cortex in left orbitofrontal and anterior cingulate cortex in first-episode psychosis patients. ³⁷ While the orbitofrontal cortex is suggested to be involved in the value representations of an outcome, the anterior cingulate cortex is suggested to play a role in the computation of the effort needed to obtain an outcome.^32-34 Takayanagi et al ³⁸ did not find any significant associations between severity of apathy and cortical thickness in a group of schizophrenia patients. However, power to detect associations may have been limited given that the study only included 15 patients, all diagnosed with DS. Farrow et al ³⁹ investigated correlations of the anterior cingulate cortex and dorsolateral prefrontal cortex gray matter volume and motor activity over a period of 20 hours, which may serve as a more objective measure for avolition and goal directed behavior. Motor activity, which was correlated with patients avolition score, was positively correlated with left anterior cingulate cortex volume. Roth et al ³⁶ found that increased severity of apathy was associated with smaller right ventral striatum volume.

With relevance to the findings of gray matter abnormalities in frontal lobe structures, Ohtani et al ⁴⁰ found that smaller FA of white matter tracts connecting the medial OFC and the rostral anterior cingulate cortex was associated with higher score of negative symptoms within the avolition-apathy domain. Hovington et al ⁴¹ did not find any group difference in FA values when comparing PNS and non-PNS patients, but did find a significant correlation between severity of anhedonia-asociality and lower FA values in the right cingulum bundle. White matter abnormalities in the internal capsule, the left superior longitudinal fasciculus, and the anterior part of corpus callosum, have also been associated with negative symptoms within the avolition-apathy domain.^42,43

A vast amount of studies has investigated striatal activation during reward anticipation (for a meta-analysis, see Radua et al. ⁹ An fMRI study using a probabilistic reversal learning task found that patients showed reduced activations in the ventromedial prefrontal cortex and the ventral striatum compared to healthy controls and that this reduction was associated to avolition. ⁴⁴ Using an emotional memory task ⁴⁵ found altered activity in the ventral striatum to be associated to more severe anhedonia. Harvey et al ⁴⁶ found that other parts of the brain reward systems to be related to anhedonia where it was negatively correlated to activity in the orbitofrontal cortex.

Expressive Deficits

There have been very few structural neuroimaging studies specifically investigating negative symptoms within the expressive domain. Rahm et al ⁴⁷ found a significant negative correlation between stereotypic thinking and volume of the right amygdala. Ballmaier et al ⁴⁸ investigated structural abnormalities of the basal ganglia (surface-based anatomical modeling methods) in 21 schizophrenia patients who were either antipsychotic-naive, or had not received antipsychotics during the previous year, compared to healthy controls. Larger surface contractions of the right anterior putamen surface were associated with higher score of affective flattening.

More studies on the topic have been performed using functional neuroimaging. Several studies have shown that patients with schizophrenia have impaired ability to discriminate between neutral and emotional faces and that this is associated to dysfunction in different brain regions but most consistently in the amygdala. Lepage et al ⁴⁹ found that both patients and healthy controls showed increased activations in amygdala during the perception of emotional faces compared with neutral faces. They also obtained a positive association between amygdala activation and the severity of blunted affect during the perception of faces of different emotions. Similarly, Gur et al ⁵⁰ found correlations between amygdala activity and blunted affect but only for fearful faces. Lindner et al ⁵¹ used masked and unmasked fearful faces and found that the amygdala response to masked fearful faces correlated positively with degree of blunted affect while the associations were reversed for the unmasked faces.