Research paperNeural complexity as a potential translational biomarker for psychosis
Introduction
Adapting to everyday stresses requires highly complex interactions between many physiological systems and their regulatory feedback loops operating across numbers of temporal and spatial scales (Goldberger et al., 2002b, Lipsitz, 2002). These dynamic interactions are typically non-linear. Changes in input can result in apparently “unpredictable” but rule-driven changes in the output that can be analyzed by complex statistical techniques such as fractal analyses and other so-called chaos theory-based approaches (Paulus and Braff, 2003). Such approaches have been used to analyze complex phenomena, both within and beyond medicine, ranging from weather and stock market fluctuations to electroencephalogram (EEG) and heart rhythms.
Complexity of physiological systems may be reduced in many disease states and aging (Goldberger et al., 2002a, Goldberger et al., 2002b), leading to degradation of their information processing, thereby making individuals less adaptable to the demands of frequently changing environments. For example, patients with severe congestive heart failure have decreased heart rate variability characterized by low entropy (i.e., regularity or reduced information content) (Costa et al., 2002) and patients with atrial fibrillation have increased randomness in their heart rate that has high entropy (i.e., increased randomness but also reduced information content) (Costa et al., 2002, Goldberger et al., 2002b, Pincus, 1991). Relatively few studies have investigated the complexity of brain physiology in psychiatric disorders.
The characteristic disorganization and unpredictability in thinking and behavior in schizophrenia (SZ) points to an intuitive connection between psychosis and chaos (Schmid, 1991). Several investigators have examined complexity in psychotic disorders. More predictable (i.e., decreased complexity) behavior in a consecutive binary choice task has been reported in SZ (Paulus et al., 1996). Decreased nonlinear complexity has been observed during REM and wake periods while undergoing sleep polysomnography in first episode SZ patients (Keshavan et al., 2004). However, some studies showing increased complexity as well (For review see (Fernandez et al., 2013; Takahashi, 2013)). Variations across studies may be related to differences in the phase of illness studied, age, medications, and illness severity. Regional brain differences in complexity measures have not been well delineated. It is also unclear if complexity differs between different psychotic disorders. Mood ratings in bipolar disorder (BP) have less complex patterns compared to healthy persons (Gottschalk et al., 1995). On the other hand, increased EEG complexity has also been observed in bipolar mania (Bahrami et al., 2005). Few studies have directly compared brain complexity measures across diagnoses within the psychosis spectrum. Finally, it is not clear whether brain complexity measures reflect merely disease related markers, or whether they represent familial vulnerability markers. There is evidence that the dynamic complexity of brain oscillations may be heritable (Anokhin et al., 2006). This points to the value of examining brain complexity measures in non-psychotic first degree relatives of psychotic disorder probands.
Complexity is often assessed using entropy-based methods (Pincus, 1991, Richman and Moorman, 2000, Rosso et al., 2002) to quantify the regularity (orderliness) of a time series. Entropy, which is a measure of randomness, increases with the degree of irregularity, reaching its maximum in completely random systems. Physiologic output in healthy conditions usually exhibits a higher degree of entropy than output in a pathological state. However, this approach could yield contradictory results in which a high degree of entropy is also observed in pathological conditions, such as heart rate rhythm in atrial fibrillation (Goldberger et al., 2002b). Therefore, a generic approach to measuring global complexity involves considering multiple time scales in a given physical system (Costa et al., 2002, Costa et al., 2005). Subsequently, MSE has been proposed based on sample entropy (Richman and Moorman, 2000) by measuring entropy over multiple time scales inherent in a time series (Costa et al., 2002).
MSE analysis introduces the notion that neither the extremes of complete regularity nor complete randomness are complex (Costa et al., 2002), and the profiles of entropy changes across different time scales are different between regular and random signal in a variety of biological signals (Costa et al., 2005). In the case of regular signal, the entropy is low across all time scales, whereas in random signal the entropy is high in short time scale and its value decays as scale factors increase; indicating the absence of information flow across different time scales in the time series. MSE has been applied to various types of biomedical data, such as electromyograms (Istenic et al., 2010), the human gait (Costa et al., 2003) and postural sway (Costa et al., 2007), EEGs (Catarino et al., 2011, Escudero et al., 2006, Mizuno et al., 2010, Park et al., 2007, Protzner et al., 2011, Takahashi et al., 2010, Yang et al., 2013b), and resting-state fMRI signal (McDonough and Nashiro, 2014, Smith et al., 2014, Yang et al., 2015, Yang et al., 2014, Yang et al., 2013a). Collectively, changes in signal dynamics toward regularity or randomness as two ways of reduced complexity may be ubiquitous in the pathology of biologic systems.
Brain complexity in SZ has been examined using electroencephalography (Keshavan et al., 2004, Takahashi et al., 2010), magnetoencephalography (Fernandez et al., 2011), and gyral folding using structural MRI (Narr et al., 2004). A valuable approach to investigating the complexity of brain activity in SZ is resting-state fMRI. Sokunbi and colleagues (Sokunbi et al., 2014) showed that SZ had increased randomness of BOLD activity as indexed by single-scale sample entropy and the Hurst exponent, suggesting that brain complexity may be increased in patients with SZ. We have recently examined the complexity of resting-state fMRI signal in an independent sample of SZ patients (Yang et al., 2015) using multiscale entropy (MSE) (Costa et al., 2002, Costa et al., 2005), and have shown reduced MSE complexity toward either regular or random patterns. The two patterns of change in complexity correlated differently with positive and negative symptoms. However, the study was limited to SZ without their relatives, and the question of diagnostic specificity was not examined.
In the current study, we applied the MSE method to assess the complexity of resting-state fMRI data from the Bipolar Schizophrenia Network for Intermediate Phenotypes (BSNIP) study. We sought to address the following questions: 1) Do probands with psychotic disorders (SZ, schizoaffective; SAD, and psychotic BP) differ from healthy controls on the MSE complexity? 2) Do MSE complexity measures differ between diagnostic categories, 3) Do MSE complexity measures correlate with dimensional measures of psychopathology and cognition? and finally 4) Do non-psychotic first degree relatives of psychotic disorder probands have alterations in brain complexity?
Section snippets
Participants
The recruitment strategy and subject characteristics of the BSNIP sample have been previously described (Tamminga et al., 2013). Patients were recruited from the community if they had a DSM IV-TR diagnosis of SZ, SAD or psychotic BP and at least one first-degree relative between the ages of 15–65 willing to participate in the study. Healthy controls were recruited from the same local communities as where the patients were recruited. Diagnoses were determined using the Structured Clinical
Demographic data and cognitive assessment
Table 1 shows the demographic and cognitive data of probands and HCs. There was no significant between-group difference in age, handedness, age of onset, duration of illness, and socioeconomic status by Hollingshead score. SZ patients had a significantly lower ratio of females than other groups (X2 =28.7; P<0.001). Between-group differences in probands were found in, YMRS (F =3.65; P=0.027), MADRS (F =6.4; P=0.002), PANSS (total scores; F =21.7; P<0.001), chlorpromazine (CPZ) equivalent dose (F
Discussion
With MSE analysis, we quantified pathologic processes of resting-state fMRI signal that exhibited either increased regularity or increased randomness. The key finding emerging from this study is that psychotic probands (e.g., psychotic BP, SAD, and SZ) showed either decreased complexity toward regularity or randomness in various brain regions. Essentially, psychotic probands differentially shared pathologically decreased complexity of resting-state fMRI signal toward randomness in the vmPFC,
Conclusion
Our observations of the overlap in brain complexity patterns between DSM IV psychotic disorders were not surprising, consistent with findings with several other biomarker investigations (Tamminga et al., 2013). Recent investigations by our group using taxometric approaches to biomarker data applies agnostic to DSM categories have revealed Biotypes that did not map on to clinical diagnoses; rather the distinctiveness of these Biotypes was supported by external validating criteria such as brain
Contributors
BH analyzed and interpreted the data, and drafted the manuscript. ACY conceptualized the study, analyzed and interpreted the data, drafted the manuscript, and provided the analysis tools. RB, BC, GDP, JAS, CT critically revised the manuscript. MK conceptualized the study, interpreted the data, and drafted the manuscript. BC, GDP, JAS, CT, and MK collected the data and are the Principal Investigators of BSNIP project.
Role of the Funding source
The funding agency had no role in the drafting, editing, or publishing of this manuscript.
Acknowledgements
This work was supported by National Institute of Mental Health (NIMH) grants MH-077851, MH-078113, MH-077945, MH-077852, and MH-077862, and the Ministry of Science and Technology (MOST) of Taiwan (grant 104-2314-B-075 -078 -MY2; 104-2745-B-075 -002).
References (86)
- et al.
Genetic influences on dynamic complexity of brain oscillations
Neurosci. Lett.
(2006) - et al.
Dorsolateral prefrontal contributions to human working memory
Cortex J. Devoted Study Nerv. Syst. Behav.
(2013) - et al.
Altered resting state complexity in schizophrenia
NeuroImage
(2012) - et al.
Atypical EEG complexity in autism spectrum conditions: a multiscale entropy analysis
Clin. Neurophysiol.
(2011) - et al.
Multiscale entropy analysis of human gait dynamics
Physica A
(2003) - et al.
Abnormal ventral frontal response during performance of an affective go/no go task in patients with mania
Biol. Psychiatry
(2004) - et al.
Complexity and schizophrenia
Prog. Neuro-Psychopharmacol. Biol. Psychiatry
(2013) - et al.
Lempel-Ziv complexity in schizophrenia: a MEG study
Clin. Neurophysiol.
(2011) - et al.
What is physiologic complexity and how does it change with aging and disease?
Neurobiol. Aging
(2002) - et al.
The Brief Assessment of Cognition in Schizophrenia: reliability, sensitivity, and comparison with a standard neurocognitive battery
Schizophr. Res.
(2004)
Norms and standardization of the Brief Assessment of Cognition in Schizophrenia (BACS)
Schizophr. Res.
Decreased nonlinear complexity and chaos during sleep in first episode schizophrenia: a preliminary report
Schizophr. Res.
Is aberrant functional connectivity a psychosis endophenotype? A resting state functional magnetic resonance imaging study
Biol. Psychiatry
Subcortical and ventral prefrontal cortical neural responses to facial expressions distinguish patients with bipolar disorder and major depression
Biol. Psychiatry
Common genetic determinants of schizophrenia and bipolar disorder in Swedish families: a population-based study
Lancet
Resting state functional connectivity of five neural networks in bipolar disorder and schizophrenia
J. Affect. Disord.
Differences in resting-state functional magnetic resonance imaging functional network connectivity between schizophrenia and psychotic bipolar probands and their unaffected first-degree relatives
Biol. Psychiatry
Assessment of EEG dynamical complexity in Alzheimer's disease using multiscale entropy
Clin. Neurophysiol.
Abnormal gyral complexity in first-episode schizophrenia
Biol. Psychiatry
Chaos and schizophrenia: does the method fit the madness?
Biol. Psychiatry
Brain signal variability relates to stability of behavior after recovery from diffuse Brain injury
NeuroImage
Neurocognitive measures of prefrontal cortical dysfunction in schizophrenia
Schizophr. Res.
Brain electrical activity analysis using wavelet-based informational tools
Physica A
Automatic denoising of functional MRI data: combining independent component analysis and hierarchical fusion of classifiers
NeuroImage
Necessity of noise in physiology and medicine
Comput. Methods Prog. Biomed.
Resting state fMRI entropy probes complexity of brain activity in adults with ADHD
Psychiatry Res.
Complexity of spontaneous brain activity in mental disorders
Prog. Neuro-Psychopharmacol. Biol. Psychiatry
Antipsychotics reverse abnormal EEG complexity in drug-naive schizophrenia: a multiscale entropy analysis
NeuroImage
Age-related variation in EEG complexity to photic stimulation: a multiscale entropy analysis
Clin. Neurophysiol.
The WU-Minn Human Connectome project: an overview
NeuroImage
Complexity of spontaneous BOLD activity in default mode network is correlated with cognitive function in normal male elderly: a multiscale entropy analysis
Neurobiol. Aging
Is mental illness complex? From behavior to brain
Prog. Neuro-Psychopharmacol. Biol. Psychiatry
Cognitive and neuropsychiatric correlates of EEG dynamic complexity in patients with Alzheimer's disease
Prog. Neuro-Psychopharmacol. Biol. Psychiatry
Brain complexity increases in mania
Neuroreport
Lateral prefrontal cortex is organized into parallel dorsal and ventral streams along the rostro-caudal axis
Cereb. Cortex
Cognitive functioning in schizophrenia, schizoaffective disorder and affective psychoses: meta-analytic study
Br. J. Psychiatry J. Ment. Sci.
EEG complexity as a biomarker for autism spectrum disorder risk
BMC Med.
DPARSF: a MATLAB toolbox for "pipeline" data analysis of resting-state fMRI
Front. Syst. Neurosci.
Identification of distinct psychosis biotypes using brain-based biomarkers
Am. J. Psychiatry
Multiscale entropy analysis of complex physiologic time series
Phys. Rev. Lett.
Multiscale entropy analysis of biological signals
Phys. Rev.
Noise and poise: enhancement of postural complexity in the elderly with a stochastic-resonance-based therapy
Europhys. Lett.
The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism
Mol. Psychiatry
Cited by (0)
- 1
These authors contributed equally to this work.