Keywords

1 Introduction

Enduring negative symptoms in patients with chronic schizophrenia correlate with decreased functional outcome (Tandon et al. 2000; Buchanan 2007; Green 2006; Kirkpatrick et al. 2006) and are refractory to common pharmacological interventions (Buchanan 2007; Buchanan et al. 2007; Galderisi and Maj 2009). It has long been recognized that negative symptoms can be secondary to medication effects, co-morbid mood disorders, or even resulting from significant cognitive disorganization. The development of the Schedule for the Deficit Syndrome (SDS) (Kirkpatrick et al. 1989) has allowed the definition of a subgroup of schizophrenia patients dominated by clusters of primary and enduring negative symptoms (Kirkpatrick et al. 1989). This subgroup of patients (deficit schizophrenia, DS) arguably suffers from the most severe and costly form of all of the subgroups of schizophrenia (Galderisi and Maj 2009; Kirkpatrick and Galderisi 2008).

Functional or anatomical probing of individual psychiatric symptoms or symptom clusters within a psychiatric syndrome is a relatively recent endeavor. Schizophrenia symptomatology varies widely affecting the cognitive, affective, and reality testing domains. Three sets of symptoms received particular attention: positive, negative, and cognition-related symptoms. Based on the fundamental differences and likely brain structures involved in the generation of these symptom clusters, an assumption can be made that the underlying physiopathologies mediating these symptoms are different. Neuropathological findings associated with schizophrenia in general have implicated most of brain structures but with significant emphasis on frontal, temporal, and thalamic regions. While the complete understanding of the pathophysiology of the DS is not possible with the current level of knowledge, most studies suggest pervasive bottom-up deficiencies that may lead to cascading information processing problems. This tentative conclusion is based on the observation that information processing abnormalities gleaned from evoked potentials (EPs) studies tend to occur earlier in the sequence of EPs and starting during periods that are commonly considered as “pre-attentive.”

Significant efforts are now being expended in attempting to ascertain whether DS is a subgroup of schizophrenia with identified abnormalities proving to be only quantitative and not qualitatively different from rest of schizophrenia subgroups. On the other hand, it is suggested that if identified abnormalities closely associated with the DS are qualitatively different then it would be possible to postulate DS as a separate psychopathological entity. The clear demarcation of the DS and its early identification could lead to better treatment and rehabilitation efforts.

Electrophysiological investigations, including electroencephalography (EEG), magnetoencephalography (MEG), evoked potentials (EPs), and sleep studies represent investigative methodology that can assess chronology at a millisecond by millisecond level of the information processing cascade in the intact behaving human. Research utilizing this methodology has pointed out to a number of highly replicable physiological aberrations in individuals suffering from schizophrenia. None the less, no biological abnormality has proven to be diagnostic or even present in a significant majority of patients. One reason commonly advanced for such is the agreed upon heterogeneity of the syndrome. By identifying the aberrations that are more specifically linked to a particular symptom (e.g., hallucination) or to a symptom cluster (positive, negative or cognitive), advances can be made toward better understanding of the neural circuitries subserving different symptom clusters, as well as the disorder as a whole. Defining the electrophysiological changes most closely linked to negative symptoms may allow guided research to probe specific negative symptoms like avolition and emotional blunting.

Physiological probing of brain function and dysfunction uses two fundamentally different technologies; functional (as contrasted to structural) neuroimaging and electrophysiology. While neuroimaging has superior spatial resolution, electrophysiology enjoys a superior temporal resolution. It is widely agreed upon that the two methodologies are complementary (Boutros et al. 2011) and efforts are underway for the simultaneous recording of both kinds of brain activity. In this review, we focus on electrophysiological investigations probing negative symptoms or the deficit syndrome and attempt to leverage the existing electrophysiology literature in schizophrenia in an attempt to investigate the possibility of a unifying hypothesis for the pathophysiology of negative symptom in schizophrenia patients. A testable hypothesis is whether positive symptoms represent a more localized dysfunction in the frontotemporal-thalamo-striatal-cortical circuitry while negative symptoms (and more fully the DS) may be secondary to a more widespread cortical-thalamo-striatal-cortical dysfunction. It is also quite possible that the nature of involvement of the frontal cortex is different in groups of schizophrenia patients with predominance of negative, positive, or cognitive dysfunctions.

2 EEG Changes in Association with Negative Symptoms

EEG abnormalities in schizophrenia have been noted since the early days of electroencephalography. The emergence of the ability to analyze EEG signals with the aid of the computer allowed the intense and detailed interrogation of the rich electrophysiological data. An increase of slow activity in the delta and theta range (1–8 Hz) was consistently reported in schizophrenia (Galderisi and Maj 2009). Spectral EEG studies have linked increased slow wave activity to the prevalence of negative symptoms (Boutros et al. 2008a).

Table 1 lists studies that directly addressed the correlation between negative symptoms and/or the deficit syndrome EEG changes. Table 1 also highlights the varying EEG methodologies used in probing negative symptoms (and schizophrenia in general) (Boutros et al. 2008a). Totally, 5 of the 11 studies using spectral analysis point to an increased slow activity (mainly theta rhythms of 4–8 Hz frequency) in association with negative symptoms. A relatively recent study supported the observation of a significant correlation between increased theta activity and negative symptoms (Venables et al. 2009). In this study the strongest correlations were obtained from frontal, central, and occipital regions while correlation between EEG data obtained from temporal regions and negative symptoms barely reached significance. This observation suggests that the increased theta seen in association with negative symptoms may be more represented in certain brain regions and may have significant implications to the eventual identification of the specific circuitry underlying the development of these symptoms (Venables et al. 2009). These findings also underline the possibility that the pathophysiology associated with negative symptoms may be widespread. The computational role of theta oscillation has been linked to periodic modulation of synaptic transmission and plasticity (Lengyel et al. 2005). Persistent theta neural activity lasting for seconds after transient stimulation has been observed in several brain regions. This activity has been taken to be indicative of the integration of inputs on long time scales (Huhn et al. 2005).

Table 1 EEG studies examining correlation with negative symptoms
Full size table

As far as delta band activities are concerned, some studies have found enhanced delta band activity in the prefrontal cortex in neuroleptic-naïve subjects with schizophrenia (Pascual-Marqui et al. 1999) or a generalized increase in delta band activity, most prominently in the anterior cingulate gyrus and temporal lobes (Mientus et al. 2002). Itoh et al. (2011) conducted the first study to assess, in first episode drug-naïve patients, if delta band activity would be increased in brain areas with a relevant relationship with the pathophysiology of schizophrenia and its correlation with the symptomatology of schizophrenia (e.g., positive and negative symptoms). Low Resolution Brain Electromagnetic Tomography (LORETA) was used to generate current source density images of delta, theta, alpha, and beta activities and localize the difference in EEG activity between patients and healthy controls. Current density for delta band activity was greater for patients in the left inferior temporal gyrus, right middle frontal gyrus, right superior frontal gyrus, right inferior frontal gyrus, and right parahippocampal gyrus. Current density values for delta band at these brain regions showed a negative correlation with the Scale for Assessment of Negative Symptoms (SANS) total score and no significant correlation with Scale for Assessment of Positive Symptoms (SAPS) score. The results of this study indicated that increased slow activity in frontal regions is not associated with positive symptoms and is more marked in subjects with fewer negative symptoms. In the light of recent conceptualization of increased delta band over frontal regions as an electrophysiological sign of reduced attention to the external input, the authors hypothesized that subjects with increased frontal delta may be able to reduce external input, cope better with positive symptoms and have less secondary negative symptoms. It is thus clear that further research is needed to more fully understand the interrelationship between increased slow wave activity, particularly in the frontal regions, and schizophrenia symptomatology.

A less consistently reported abnormality is altered reactivity and/or predominance of faster rhythms (beta activity). Noh et al. (2013) analyzed magnetoencephalography (MEG) data in chronic patients with schizophrenia and found reduced gamma power and increased beta band synchronization. Using an index of intra and interregional cortical functional connectivity (the amount of coupled local and global feedback (CLGF) circuits) they found that dysfunctional connectivity was related to the abnormal synchronization of the beta band and to negative symptoms.

Another important approach of describing dynamical systems is nonlinear analysis, called chaos theory, which allows researchers to determine changes in the dimensionality of the system over time (Kotini and Anninos 2002). A higher rate of phase-state transitions (i.e., higher dimensionality) characterizes more complex systems. Earlier work revealed that nonlinearity scores were significantly lower during awake state in schizophrenia patients compared to control subjects suggesting that there may be diminished interplay between different generators of the various EEG rhythms (Keshavan et al. 2004). Decreased nonlinear complexity correlated with neurocognitive deficits. Other groups found an increase in the dimensional complexity in schizophrenia patients (Koukkou et al. 1993; Rockstroh et al. 1997). Kotini and Anninos (2002) when using MEG to examine nonlinearity in schizophrenia patients also reported lower dimensional complexity. One possible contributor to the discrepancy is heterogeneity of study samples. A readily testable hypothesis would be that decreased dimensionality would correlate with increased cognitive deficit and possibly negative symptoms.

Yet another recently evolving methodology is to stimulate the cortex using transcranial magnetic stimulation (TMS) and examine the resultant EEG effects. While yet to be applied specifically to probe the deficit syndrome, Guller et al. (2012) used this methodology to test the hypothesis that direct physiological stimulation of the cortex will produce an abnormal thalamic response in individuals with schizophrenia. They stimulated the precentral gyrus with single pulse TMS (spTMS) and measured the response to this pulse in synaptically connected regions (thalamus, medial superior frontal cortex, insula) using concurrent functional magnetic resonance imaging. Compared with healthy subjects, patients with schizophrenia showed a reduced response to spTMS in the thalamus, medial superior frontal cortex, and in the insula. Functional connectivity analyses revealed weaker thalamus–medial superior frontal cortex and thalamus–insula connectivity in patients with schizophrenia compared with control subjects. Subsequently, the same group utilized the same technology probing EEG effects to assess the natural frequency of the posterior parietal, motor, premotor, and prefrontal cortices in patients with schizophrenia and healthy control subjects. High-density EEG recordings during TMS of four cortical areas were performed. Several TMS-evoked EEG oscillation parameters, including synchronization, amplitude, and natural frequency, were compared between the schizophrenia and control groups. Patients with schizophrenia showed a slowing in the natural frequency of the frontal/prefrontal regions compared with control subjects (from an average of a 2-Hz decrease for the motor area to an almost 10-Hz decrease for the prefrontal cortex). The prefrontal natural frequency of individuals with schizophrenia was slower than in any healthy comparison subject and correlated with both positive and negative syndrome scale scores suggesting that these abnormalities may be common between patients dominated by positive or negative symptoms and further implicates thalamic abnormalities in schizophrenia. It could be speculated that frontothalamo-striatal-frontal circuit dysfunctions are common to all schizophrenia patients and perhaps contributing to the common coexistence of the varying manifestations of the disorder. According to our hypothesis advanced above, the development of a DS picture may require involvement of other cortical regions beyond frontal and perhaps temporal regions.

3 Evoked Potentials (EP) Changes Associated with Negative Symptoms

Disturbances in sensory processing have been suggested as significant contributors to the deficit state in both clinical and neurophysiological studies (Turetsky et al. 1998; Mucci et al. 2007; Li et al. 2013). Two studies have thus far demonstrated qualitative differences between DS and NDS (nondeficit) patients utilizing EP measures (Turetsky et al. 1998; Mucci et al. 2007). Thus, surrogate markers of sensory information processing are promising tools for probing the pathophysiology of the DS as well as delineating differences between DS and NDS physiological aberrations. Table 2 lists studies that directly examined the correlation between negative symptoms and evoked potential changes.

Table 2 EPs
Full size table

3.1 Mid-Latency Auditory Evoked Responses and Sensory Gating

The mid-latency auditory evoked responses (MLAERs) have been extensively used to study information processing both in mental health and disease. Numerous studies have shown two particular components to be abnormally reduced in amplitude in patients with schizophrenia. These two components are the N100 (a negative component seen approximately 100 ms following an auditory stimulus) and the P50 (a positive component seen about 50 ms after the presentation of an auditory stimulus) (Erwin et al. 1991). The N100 component has been extensively examined in schizophrenia patients. The majority of studies report decreased amplitude of the N100 which is not readily attributable to medication effects (Rosburg et al. 2008). N100 amplitude is commonly (but not invariably) found to be reduced in patients with schizophrenia (Rosburg et al. 2008) and in unaffected first-degree relatives of the same patients (Ahveninen et al. 2006). The relationship between N100 amplitude and specific schizophrenia symptoms remains unclear (Rosburg et al. 2008).

The P50 MLAER has also been used extensively to examine the phenomenon of amplitude attenuation with stimulus repetition. The term “sensory gating” has been linked with studies of sensory inhibition utilizing the P50 MLAER in a paired stimulus paradigm (PSP). Abnormal sensory gating has been proposed as a fundamental mechanism by which psychotic symptoms evolve (Freedman et al. 1991). In a PSP, two identical stimuli (S1 and S2) are delivered with short interstimulus interval of 500 ms and a longer interpair interval of 8–10 s (Adler et al. 1982). A sensory gating deficit has been repeatedly demonstrated in schizophrenia patients (Adler et al. 1982; Boutros et al. 1991). Meta-analysis of the P50 gating deficit in groups of nonselective schizophrenia, patients found the effect size to be more modest than earlier reports suggested (Patterson et al. 2008). Heterogeneity of both methodology and composition of patient groups were suggested as possible causes for variation in results. A number of studies suggest that the gating deficit, particularly of the N100 component, may be more strongly found in association with negative symptoms (Boutros et al. 2009).

The gating deficit as assessed by the P50 component was examined in relationship to negative symptoms and was not found to correlate with any symptom cluster in earlier studies (Adler et al. 1990; Boutros et al. 2004). Subsequent investigations of gating of the P50 MLAER component reported a significant correlation with negative symptoms (Ringel et al. 2004; Louchart-de la Chapelle et al. 2005; Arnfred 2006). More recently Santos et al. (2010) investigated P50 gating in patients with DS and in those with nondeficit schizophrenia (NDS). These authors did not find differences in P50 gating between the two groups of patients, but found an association between the abnormality of P50 gating and poor functional outcome. These contradictory findings suggest that the relationship between primary, enduring negative symptoms and the P50 gating deficit are not simple or straightforward and may depend strongly on the specific composition of the study sample. Given the documented high prevalence of smoking among schizophrenia patients it is also important to control for smoking status of the patient and the time elapsed between the last cigarette and P50 testing as transient improvement in P50 gating following nicotine/cigarette smoking has been demonstrated (Leonard et al. 2007). It is also possible that the effect size of the observation is not robust enough to be detected in smaller samples or in the presence of a high noise to signal ratio as is common in P50 studies due to the relative small amplitude of the component (Boutros et al. 2009). Furthermore, the specific scale or instrument used to assess the negative symptoms (or for that matter any symptom cluster) may also influence the correlations identified (Boutros et al. 2009). An alternative hypothesis is that a mediating variable (not as yet identified), cross-correlated with negative symptoms and gating deficits (e.g., poor outcome), is responsible for the association between P50 gating deficit and negative symptoms. Further studies are needed to clarify the issue.

Evidence for a frontal cortex role in mediating both auditory and somatosensory habituation has been provided by MEG studies (Bowyer et al. 2007; Weiland et al. 2008), EEG studies (Korzyukov et al. 2007; Garcia-Rill et al. 2008) and from direct brain recordings (Grunwald et al. 2003; Boutros et al. 2008a, b). Physiological data have indicated that the inability to suppress irrelevant inputs, coupled with difficulty in novelty detection (which may be related and indexed here by the MMN), impairs the coding at the beginning and ending of discrete events. Information is stored with incorrect spatiotemporal tags (Knight et al. 1995). A number of studies documented a working memory deficit in schizophrenia patients (Goldman-Rakic 1999). It has further been postulated that it is the inability to gate irrelevant information during the recall period, that causes the deficit to become manifest. Chao and Knight (1995) tested this hypothesis in a number of patients with dorsolateral prefrontal cortex (DLPFC), temporal–parietal junction (TPJ), or posterior hippocampus lesions. They showed that patients with prefrontal lesions were significantly impaired by distractors at all delays while patients with temporo-parietal lesions performed similar to control subjects. Individuals with DLPFC lesions exhibited increased amplitudes of the mid-latency auditory evoked responses (MLAERs) beginning 25–35 ms following auditory stimulation (Knight et al. 1989). It is possible to speculate that the preattentive gating abnormality (mediated by the P50) may be common to all schizophrenias while the later occurring (or early attentive) gating abnormalities (reflected by either the N100 or the P200 EPs) may be more linked to negative symptoms (also a readily testable hypothesis).

Smucny et al. (2013) most recently used time–frequency analysis to compare sensory gating at the beta (15–26 Hz) and gamma (30–50 Hz) frequencies between schizophrenia patients and healthy controls. Relative to controls, patients showed impaired gating of total beta and gamma power. Poor beta gating was associated with negative symptoms. Time–frequency analysis of beta and gamma gating may thus be a translational method of assessing the genetic basis of gating deficits in schizophrenia.

3.2 Mismatch Negativity

The mismatch negativity (MMN) is an early EP (with latency around 150 ms from the stimuli onset) related to automatic probing of auditory sensory traces of a repetitious stimulus by a deviant one elicited in an oddball paradigm (Näätänen et al. 1978; Sams et al. 1985). MMN is considered a preattentive process and thus mainly a bottom-up function. While MMN has been shown to be diminished in a number of studies of schizophrenia patients, not all studies find similar deviations (some find duration and others find frequency deviation MMN alterations). We propose that one possible explanation of the variance in the findings is related to inclusion of different patient samples and not paying specific attention to negative versus positive symptoms as well as not paying attention to enduring primary negative versus secondary negative symptoms. The MMN has not yet been specifically investigated in well-characterized DS patients. Based on the assumption that DS patients suffer from an early (preattentive and bottom-up) sensory processing deficit, we fully expect that future MMN investigation will reveal abnormalities of this component in DS patients. At the time of preparing this chapter, only a relationship between diminished frequency MMN and executive dysfunction has been reported (Toyomaki et al. 2008).

3.3 The P300

The P300 response is a positive deflection appearing around 300 ms from stimuli onset in response to rare novel or deviant stimuli in an auditory oddball paradigm. Its amplitude is proportional with the amount of attentional resources allocated to the task and memory performance (Polich 2007). The P300 as well as the MMN have been found to be abnormally small (and sometimes delayed) in schizophrenia populations (Javitt et al. 1993; Näätänen and Kähkönen 2009; Jeon and Polich 2003; Gjini et al. 2010). The P300 response represents processing of information at more advanced cognitive levels, such as a shift of attention, context updating, or orienting to a relatively novel or deviant stimulus (Polich and Kok 1995). Examination of the P300 is of particular importance as evidence suggests that it may be relatively spared in DS patients (Mucci et al. 2007; Jeon and Polish 2003). It is of importance to note that a meta-analysis of the P300 ERP in schizophrenia found a correlation with positive but not negative symptoms (Jeon and Polish 2003). Mucci et al. (2007) were able to corroborate this finding in a group of well-characterized DS patients. Confirmation of this important observation would significantly contribute to the delineation of the neural circuitry implicated in DS. Engagement of frontal attention mechanisms during evaluation of novel incoming stimuli produces the P3a response with a midline centrofrontal maximum as compared to the classic P300 (i.e., P3b) activity to attended target stimuli showing a midline parietal maximum and has been related to context updating operations and subsequent memory storage (Polich 2007). A recent fMRI study (Wolf et al. 2008) showed a negative correlation of novelty-induced BOLD signal in the left frontal cortex with SANS scores, indicating lower levels of activation in patients with more severe negative symptoms. No significant correlations were identified for novel-induced BOLD signal and positive symptoms; also no significant correlations for target-induced BOLD with either positive or negative symptoms. Based on the limited available literature examining the correlations between P3 and negative symptoms we predict that DS patients will exhibit deficit in the ability to respond to novelty reflected by generating lower amplitude P3a components, but preserved target P300 (P3b).

Kim et al. (2013) analyzed four event-related potential (ERP) components (P100, N170, N250, and P300) and their source activities using EEG data acquired from 23 schizophrenia patients while they were presented with facial emotion picture stimuli. Correlations between positive and negative syndrome scale (PANSS) scores and source activations during facial emotion processing were calculated to identify the brain areas affected by symptom scores. Analysis demonstrated that PANSS positive scores were negatively correlated with major areas of the left temporal lobule for early ERP components (P100, N170) and with the right middle frontal lobule for a later component (N250), which indicates that positive symptoms affect both early face processing and facial emotion processing. On the other hand, PANSS negative scores were negatively correlated with several clustered regions, including the left fusiform gyrus (at P100), most of which are not overlapped with regions showing correlations with PANSS positive scores. These recent results suggest that positive and negative symptoms affect independent brain regions during facial emotion processing.

3.4 Error-Related and Correct-Response Negativity (ERN and CRN)

Deficits in self monitoring are a core feature of cognitive dysfunction in schizophrenia, and may be the basis for disturbances of self and lack of insight, ultimately impacting social functioning. DS patients are characterized by a significant lack of insight. However, the functional and structural neural correlates of such deficits in self monitoring are not well understood (Araki et al. 2013). This group of researchers investigated this issue using measurements of neurophysiological and structural brain indices, i.e., error-related and correct-response negativity (ERN and CRN) ERPs, and gray matter volume of the anterior cingulate cortex (ACC), and tested whether the association between these indices is altered in patients with schizophrenia compared to healthy controls. The two groups did not differ in ERN amplitude. In contrast, schizophrenia patients showed significantly larger CRN amplitudes than did healthy subjects. Although the two groups did not significantly differ in gray matter volume of the ACC subregions, a significant negative correlation was found between ERN amplitudes at the frontocentral electrodes and absolute gray matter volumes of the left region of ACC only in healthy controls. These results suggest a disruption of function–structure coupling of the brain regions subserving self monitoring in schizophrenia. A direct comparison between DS and non-DS schizophrenia patients would be informative in this regards.

3.5 Anticipatory Components: Contingent Negative Variation (CNV) and Stimulus-Preceding Negativity (SPN)

The negative symptoms construct has been refined with the modern conceptualization identifying 5 core component that have been shown to separate into two domains: 1) a motivational dimension consisting of anhedonia, avolition, and asocialty and 2) a diminished expressivity dimension consisting of restricted affect and alogia (Strauss et al. 2013). In schizophrenia reward deficits involve more than just anhedonia (Foussias and Remington 2010). In some cases, while the ability to experience pleasure may even remain intact, patients with schizophrenia more consistently demonstrate an impaired anticipation of reward, not immediately available (Gold et al. 2008) and impairments in several other facets of reward processing involved in motivation (Strauss et al. 2013).

ERP components studied in relation to anticipation of reward include the contingent negative variation (CNV), a slow negative brain wave shown to reflect the anticipation of or orienting to the upcoming stimulus and response preparation, and has been related to preparatory attention, motivation, and response readiness (Walter et al. 1964), and the stimulus-preceding negativity (SPN), a negative ERP detected in paradigms that involve anticipation of feedback about the correctness of prior performance or stimuli that are motivationally significant (van Boxtel and Bocker 2004). Schevernels et al. (2014) showed that reward anticipation was linked to early cue processing components, as well as the early and later parts of the CNV. Wynn et al. (2010) investigated anticipatory deficits in subjects with schizophrenia in a paradigm involving a cued motor response (CNV) and no motor response (SPN) and the relationships of these ERP components abnormalities with self-reported trait anhedonia or anticipatory pleasure and clinically rated negative symptoms. Patients demonstrated generally lower CNV and SPN across pleasant, neutral, and unpleasant conditions; SANS total score did not correlate with the ERP variables; higher trait anhedonia was related at a trend-level to lower overall SPN.

Other facets of motivation found to be impaired in schizophrenia include difficulties using internal representations of emotional experiences, previous rewards, and motivational goals to drive current and future behavior, deficits that have major clinical significance in terms of functional capacity (Barch and Dowd 2010). The feedback-related negativity (FRN) is an ERP component that has been localized to the ACC (Nieuwenhuis et al. 2005) and has been hypothesized to reflect the function of a performance monitoring/evaluative system that rapidly assesses the motivational impact and/or salience of environmental feedback (Segalowitz et al. 2010). Fitting this view, the FRN has been found to be sensitive to both errors in reward prediction, negative valence of emotional stimuli (Pfabigan et al. 2011) and the magnitude of outcome (Holroyd and Krigolson 2007; Holroyd et al. 2004). In their study, Horan et al. (2012) tried to clarify the scope of feedback processing impairments in schizophrenia analyzing FRN during a simple monetary gambling task. The authors also investigated the relationships between FRN abnormalities and negative symptoms, assessed using the BPRS. They showed that patients and controls demonstrated comparable FRN differentiation between reward and nonreward feedback and higher positive symptoms were associated with greater differences between FRN to positive and negative feedback. There were no significant or trend-level correlations for negative symptoms. In their opinion the use of the BPRS may have limited ability to detect an association between negative symptoms and reward processing because the BPRS negative symptom subscale focuses on expressive symptoms (e.g., blunted affect) whereas experience-related symptoms (e.g., avolition, asociality) have a stronger theoretical link to feedback and reward processing (Blanchard et al. 2011).

3.6 Loudness Dependence of Auditory Evoked Potentials (LDAEP)

There is an evidence that alterations of serotonin (5-HT) system functioning also play a crucial role in the pathophysiology of disabling negative symptoms. From post mortem and genetic studies on patients with negative symptoms a 5-HT dysfunction is documented. In addition, atypical neuroleptics and some antidepressants improve negative symptoms via serotonergic action. So far, no research has been done to directly clarify the association between the serotonergic functioning and the extent of negative symptoms. Wysse et al. (2013) based on the above, examined the status of brain 5-HT level in negative symptoms in schizophrenia by means of the loudness dependence of auditory evoked potentials (LDAEP). The LDAEP provides a well-established and noninvasive in vivo marker of the central 5-HT activity. They investigated 13 patients with schizophrenia with predominant negative symptoms treated with atypical neuroleptics and 13 healthy controls. The LDAEP of the N1/P2 component was evaluated by dipole source analysis and single electrode estimation at Cz. Psychopathological parameters, nicotine use, and medication were assessed to control for additional influencing factors. Schizophrenia patients showed significantly higher LDAEP in both hemispheres than controls. Furthermore, the LDAEP in the right hemisphere in patients was related to higher scores in scales assessing negative symptoms. A relationship with positive symptoms was not found. These data might suggest a diminished central serotonergic neurotransmission in patients with predominant negative symptoms.

It should be noted that only three papers specifically addressed ERPs in the deficit syndrome (Turetsky et al. 1998; Mucci et al. 2007; Li et al. 2013). Two studies provided data suggestive that the deficit syndrome is not simply a severe form of schizophrenia but more likely a separate clinical entity. Turetsky et al. (1998) while not employing the SDS, used the BPRS, SANS, and SAPS to address the criteria for the deficit syndrome. They examined the P300 component and found two patterns that cannot be seen as different grades of the same process. Patients with NDS showed the greatest reduction over the left temporal regions while the deficit subgroup showed the greatest reduction over the right parietal region. The second study specifically addressing DS, using the SDS, found a double dissociation where only NDS patients exhibited the asymmetrical left temporal P300 amplitude deficiency, while DS patients exhibited a decreased amplitude of the N100 (Mucci et al. 2007). Most recently, Li et al. (2013) compared a number of EPs between 21 patients with DS, 38 patients with NDS schizophrenia and 50 healthy control subjects. They included the P300 (both P3a and P3b), MMN, amplitudes and gating of the P50 MLAER, and the contingent negative variation (CNV). To our knowledge this is the only report examining the CNV in relationship to negative symptoms schizophrenia. The authors reported a great deal of similarity between the two schizophrenia groups. As compared to the healthy controls both patient groups exhibited delayed N1, N2, MMN, and P3a latencies as well as reduced N1, N2, and CNV amplitudes. Similarly, P50 gating was decreased in both groups. Only NDS, when compared to controls, showed delayed latency of P3b. Only DS patients showed delayed point A in CNV which correlated with poorer Global Assessment of Functioning scale but not with any individual negative symptom. These recent data suggest a significant overlap between the DS and NDS syndromes but with some distinctive characteristics that need further explorations in larger and perhaps unmedicated patients (Li et al. 2013).

4 Sleep Changes in Association with Negative Symptoms

Table 3 lists studies that directly reported on the association between sleep architectural changes and negative schizophrenia symptoms.

Table 3 Sleep
Full size table

Decreased delta sleep in association with negative symptoms has been a relatively consistent finding in schizophrenia patients (Tandon et al. 2000). Of the 13 studies identified, 10 reported decreased slow wave sleep (SWS) in association with negative symptoms. Of these 10 papers, seven found a significant negative correlation between the severity of negative symptoms and percent of SWS (i.e., with increased severity of negative symptoms, less SWS is noted).

Recent sleep studies depend heavily on computer-based analysis. Sekimoto et al. (2007) utilized a specialized program (the Medilog Sleep Analyzing Computer) to perform period-amplitude analysis of the delta wave count (delta half-wave analysis). They reported that schizophrenia patients in general exhibited a significant decrease in the percent of stage 2 sleep and a marked decrease of SWS. The delta half-wave count in the bilateral frontal regions was inversely correlated with BPRS negative symptoms scores. Poulin et al. (2008) reported a negative correlation between sleep absolute alpha activity and negative symptoms scores. Sarkar et al. (2010) reported a significant decrease in SWS, stages one and two as well as total sleep time in general but with no correlation with negative symptoms. On the other hand, Yetkin et al. (2011) found no evidence of a SWS abnormality in a group of 13 male schizophrenia patients (undifferentiated type).

Relatively more recently, Sekimoto et al. (2011) reported a significant inverse correlation between negative symptoms scores and delta wave counts in all regions examined (frontal, central, parietal, occipital). Utilizing self-report (Pittsburgh Sleep Quality Index), Lunsford-Avery et al. (2013) observed a strong association between negative symptoms severity (and not positive symptoms) and increased sleep dysfunction (no PSG data obtained).

5 Conclusions

Progress in the investigation of negative symptoms pathophysiology has been hindered by intrinsic heterogeneity of this symptom cluster. Primary and enduring negative symptoms are recognized as distinct disease processes with respect to broadly defined negative symptoms. Recent conceptualizations and factor analytic studies of both primary and broadly defined negative symptoms support two distinct domains: avolition and impaired emotional expression, probably related to separate pathophysiological mechanisms (Cohen et al. 2007). These sources of heterogeneity have seldom been considered in the electrophysiological literature. However, the field is rapidly changing and studies of the deficit syndrome as well as investigations of the separate domains of negative symptoms are rapidly increasing.

Two prominent/reasonably consistent electrophysiological correlates of negative symptoms emerge through this review: increased slow frequencies during wakefulness (as assessed by awake spectral EEG) and decreased slow wave sleep during night time recording.

While the largest group of studies identified were those related to EP/ERP findings, a consistent or strong trend was difficult to identify perhaps due to the tendency of different research groups to examine one particular EP component like P300, MMN, or sensory gating. Studies examining a number of EP components simultaneously are sparse (Li et al. 2013). Despite the small number of studies and the varying EPs examined, the literature points to a deficit in the sensory gating of the mid-latency evoked responses reported from more than one laboratory (Boutros et al. 2009; Ringel et al. 2004; Louchart-de la Chapelle et al. 2005). Furthermore, a decreased amplitude of the N100 response has also been linked to the deficit syndrome (Mucci et al. 2007). It is of importance to note that a meta-analysis of the P300 ERP in schizophrenia found a correlation with positive but not negative symptoms (Jeon and Polich 2003). Mucci et al. (2007) were able to corroborate this finding in a group of well-characterized DS patients.

As is a direct reflection of the research expertise represented in a particular laboratory, most research groups tend to probe one physiological variable (e.g., ERP, EEG, or sleep). Multicomponent electrophysiological studies will help further define the abnormalities detected. For example, if an EEG or EP abnormality disappears during sleep this would influence the understanding of its pathophysiology as compared to the abnormality being persistent during sleep. If an EP abnormality is linked to a particular EEG state, this finding would be of significance for the identification of the neural circuitry mediating this abnormality.

The current level of knowledge is inadequate to propose a unifying theory of the pathophysiology underlying the described anomalies associated with the DS. However, most studies suggest pervasive bottom-up deficiencies that may lead to cascading information processing problems. This tentative conclusion is based on the observation that information processing abnormalities gleaned from EP studies (the only investigative methodology that can assess chronology at a millisecond by millisecond level in the intact behaving human), tend to occur earlier in the sequence of EPs and starting during periods that are commonly considered as “pre-attentive.” Sleep deviations suggest a serious abnormality of the restorative deep sleep stages and the awake EEG abnormality suggest difficulty generating the faster frequencies, reflecting decreased ability to generate efficient smaller neuronal ensembles to deal with more focused or effective information processing.

A unifying hypothesis of the deviant oscillations during wakefulness (increased slow activity) and decreased slow wave sleep (SWS), could point to a dysfunction involving thalamocortical circuitry (Pinault 2011; Sekimoto et al. 2011). Thalamocortical circuits exhibit two fundamentally different modes of operation across the sleep-wake cycle: a state of tonic activation (desynchrony) during waking and REM sleep and a state of rhythmic synchronized activity during SWS (Steriade and Llinas 1988). Thalamic relay receives significant input from a number of brain stem structures and thus are subject to changes with a number of ascending neurotransmitter inputs. Furthermore, thalamic relay neurons also send collateral projections to the thalamic reticular nucleus which are reciprocally inhibitory (GABAergic) with thalamic nuclei (Siegel 2011). The thalamus (and more specifically the reticular nucleus) has been proposed to be important for the function of sensory gating which has been repeatedly shown to be deficient in schizophrenia and may be more specifically associated with negative symptoms (Krause et al. 2003). Kirkpatrick and Buchanan (1990) proposed a neural circuit that may be at the heart of the DS. Components of this circuit include the amygdala, periamygdalar cortex, and parts of the prefrontal cortex. A number of thalamic nuclei (including the anterior, midline, mediodorsal, lateral anterior, and lateral dorsal as well as the intralaminar) have extensive connections with all these components (Clarke et al. 2010). The involvement of the thalamus is also supported by observation of worsening of somatosensory gating in patients with thalamic strokes with recovery of the function over time (Stain et al. 2002). Recent reviews suggest a central role for thalamic abnormalities in the generation of schizophrenia symptomatology (Byne and Hazlett 2009). Recent studies suggested an association between DS and genuine movement disorders (unrelated to antipsychotic treatment) pointing to cortico-striatal-thalamic circuits (Peralta et al. 2014). It could be hypothesized that abnormalities of the latter circuits might be related to avolition, while those concerning the limbic-thalamic circuits to the emotional expression domain of the negative symptoms.

Based on the above, it can be stated that while research on the electrophysiological correlates of the deficit syndrome and enduring negative symptoms remains minimal, available data strongly support the need and likely profitability of this line of investigation. Most notably is the absence of studies where EPs, EEGs and sleep studies are performed in the same individuals in order to examine the correlation and interrelationships among these deviations.