Our metaanalysis of EEG resting state studies in schizophrenia highlight inconsistencies in methods and results as well as the non-specificity of results across disorders.

Schizophrenia is characterized by a range of psychotic symptoms such as hallucinations, delusions, reduced affect and disorganized thought, as well as a number of cognitive impairments in language, reasoning and some aspects of basic sensory processing.  Schizophrenia is therefore not the result of dysfunction in a particular region in the brain, but of disturbances across a distributed network of regions, and, in particular, the connections between them (see here for a recent review).

A number of studies have attempted to find differences in the EEG between people diagnosed with Schizophrenia and normal controls.  Here we present a short meta-analysis of 18 studies conducted over the last ~15 years that focus on changes in the power spectrum.  The majority (16) of these studies used an eyes-closed approach, although a minority (2) chose instead to study patients and healthy controls with their eyes-open.

We highlight the challenges of this approach and reveal the level of consistency (and inconsistency) within this subset of the literature.

 

Methodological Inconsistencies

The range of methodological inconsistencies span a broad range – from the way in which Schizophrenia and symptom severity is assessed to the medication profile of patients in the study to analytical methods.  Each of these can influence results and preclude easy comparison of one study to another.

Diagnostic/Screening questionnaires	DSM-III, DSM-IV; ICD-10; Positive and Negative Syndrome Scale (PANSS); BPRS, Brief Psychiatric Rating Scale
Eyes open/Closed	16 studies eyes open; 2 studies eyes closed
Medication status	All medicated (4 studies), mix of medicated and unmedicated (naive) (8 studies), unmedicated (naive or free; for a minimum of 3 days-3 months depending on the study) (6 studies)
Gender Ratio	Average: 70:30, males:females; ranges from 100:0 to 43:57
Average Age of Participants	32 years, average ranges from 26-44

 

Types of Averaging	Average, Ear lobes, Linked ears, Mastoids, Nose
EEG Recording Duration	Ranged from 2-6 minutes
Epoch length for FFT	Ranged from 1 -60 seconds
Examples of FFT parameters used.	Welch’s method; spectrogram.m (Matlab); Hanning window; Hamming window; overlap from 20-80%; taper from 10-100%

 

Resting States Results

In general, patients with schizophrenia display greater power in the lower frequency bands (delta and theta) compared to healthy controls (see table below). This is in line with other meta-analyses of the data which have concluded that “most consistent results related to the increased preponderance of slow rhythms in schizophrenia patients”. However, power increases have also been reported in the gamma band, whilst reported changes within the alpha band are highly inconsistent (e.g. both increases and decreases reported with regional variability across studies – see Alpha Alpha Everywhere for more discussion on the level of inconsistency observed in the alpha band).  The inconsistent results in the higher frequencies may be on account of methodological differences but may also reflect a wide variability within ‘Schizophrenia’ itself.

 

Number of studies (out of 18) reporting significant increases or decreases in absolute power	DELTA		THETA		ALPHA		BETA		GAMMA
Schizophrenia – eyes closed	8		10		2	5		1	5
Schizophrenia – eyes open			2		1			1

 

However, even if the inconsistencies in the pattern within the Schizophrenia studies could be resolved, more significant is that the general pattern is not unique to schizophrenia. Increases in the lower delta and theta frequency bands is also observed in patients with ADHD, OCD and to some extent depression and internet addiction (see The Brain on Games). ADHD does show more consistent decrease in the beta band (see ADHD and the Theta/Beta Ratio), relative to Schizophrenia. However, when considering the magnitude of the theta increases, which appear to be higher for Schizophrenia, the theta/beta ratios are likely to span a very similar range in both conditions.  On the other end of the spectrum, the gamma increases (which are not well studied in general) are also observed in patients with opioid addictions – see Opiate Addiction in the EEG).

Symptom severity and the power spectrum

While reporting an overall difference between groups is at best merely suggestive, how do the overall magnitude of changes relate to these reported changes?  Out of these 18 studies, 9 included an analysis which related changes within the spectral bands against symptom severity (as measured by questionnaires such as the The Positive and Negative Syndrome Scale; PANSS).  The results are mixed:

• 4 studies found no correlations between symptom severity (e.g. measured with the PANSS) and the changes within the spectral bands; [4,10,12,17]
• 2 studies found correlations between subjective measures (e.g. anergia, depression, activation; symptom severity) and changes within the gamma band [r>0.37] [1, 2]
• 1 found that level of functioning (LOF) was related to theta-alpha changes (r = -0.23) [7]
• 2 found significant correlations between the negative subscale of the PANSS and changes in the theta band [3, 18], whilst one of these also found correlations with the alpha, beta bands [3] (all r>0.31).

In general, the correlations are not particularly strong, certainly not sufficiently strong to have any predictive or diagnostic value.

Beyond the Power Spectrum

The results of our mini metaanalysis makes two key points.  First, methodological standardization in the field would go a long way towards resolving ambiguity of results. Second, the power spectrum itself, which is a very macro feature of the EEG, has limited degrees of freedom, and with similar patterns of changes across a variety of disease states, it is not a specific measure. This cautions any interpretation of results of studies done with one disease group in isolation as unique to that disease. Rather for real insight, simultaneously study multiple disorders in the context of multiple features of the EEG (see related post.

As a last note we point out that there are several other studies that have looked beyond the resting state power spectrum. These include measures of  coherence or connectivity [1] as well as changes in both spectral bands (see here for a review) and ERPs during different task paradigms (see here for a review).  We have not, however, systematically looked at results across these studies.

 

List of studies included in this meta-analysis:

1. Bandyopadhyaya, D., Nizamie, S., Pradhan, N., & Bandyopadhyaya, A. (2011). Spontaneous gamma coherence as a possible trait marker of schizophrenia—An explorative study. Asian Journal Of Psychiatry, 4(3), 172-177.
2. Baradits, M., Kakuszi, B., Bálint, S., Fullajtár, M., Mód, L., Bitter, I., & Czobor, P. (2018). Alterations in resting-state gamma activity in patients with schizophrenia: a high-density EEG study. European Archives Of Psychiatry And Clinical Neuroscience.
3. Garakh, Z., Zaytseva, Y., Kapranova, A., Fiala, O., Horacek, J., & Shmukler, A. et al. (2015). EEG correlates of a mental arithmetic task in patients with first episode schizophrenia and schizoaffective disorder. Clinical Neurophysiology, 126(11), 2090-2098.
4. Goldstein, M., Peterson, M., Sanguinetti, J., Tononi, G., & Ferrarelli, F. (2015). Topographic deficits in alpha-range resting EEG activity and steady state visual evoked responses in schizophrenia. Schizophrenia Research, 168(1-2), 145-152.
5. Hanslmayr, S., Backes, H., Straub, S., Popov, T., Langguth, B., & Hajak, G. et al. (2012). Enhanced resting-state oscillations in schizophrenia are associated with decreased synchronization during inattentional blindness. Human Brain Mapping, 34(9), 2266-2275.
6. Harris, A., Melkonian, D., Williams, L., & Gordon, E. (2006). Dynamic spectral analysis finding in first episode and chronic schizophrenia.. International Journal Of Neuroscience, 116(3), 223-246.
7. Hong, L., Summerfelt, A., Mitchell, B., O’Donnell, P., & Thaker, G. (2012). A shared low-frequency oscillatory rhythm abnormality in resting and sensory gating in schizophrenia. Clinical Neurophysiology, 123(2), 285-292.
8. Kam, J., Bolbecker, A., O’Donnell, B., Hetrick, W., & Brenner, C. (2013). Resting state EEG power and coherence abnormalities in bipolar disorder and schizophrenia. Journal Of Psychiatric Research, 47(12), 1893-1901.
9. Kim, J., Lee, Y., Han, D., Min, K., Lee, J., & Lee, K. (2015). Diagnostic utility of quantitative EEG in un-medicated schizophrenia. Neuroscience Letters, 589, 126-131.
10. Kirino, E. (2004). Correlation between P300 and EEG Rhythm in Schizophrenia. Clinical EEG And Neuroscience, 35(3), 137-146.
11. Mientus, S., Gallinat, J., Wuebben, Y., Pascual-Marqui, R., Mulert, C., & Frick, K. et al. (2002). Cortical hypoactivation during resting EEG in schizophrenics but not in depressives and schizotypal subjects as revealed by low resolution electromagnetic tomography (LORETA). Psychiatry Research: Neuroimaging, 116(1-2), 95-111.
12. Mitra, S., Nizamie, S., Goyal, N., & Tikka, S. (2015). Evaluation of resting state gamma power as a response marker in schizophrenia. Psychiatry And Clinical Neurosciences, 69(10), 630-639.
13. Ranlund, S., Nottage, J., Shaikh, M., Dutt, A., Constante, M., & Walshe, M. et al. (2014). Resting EEG in psychosis and at-risk populations — A possible endophenotype?. Schizophrenia Research, 153(1-3), 96-102.
14. Schug, R., Yang, Y., Raine, A., Han, C., Liu, J., & Li, L. (2011). Resting EEG deficits in accused murderers with schizophrenia. Psychiatry Research: Neuroimaging, 194(1), 85-94.
15. Shreekantiah Umesh, D., Tikka, S., Goyal, N., Nizamie, S., & Sinha, V. (2016). Resting state theta band source distribution and functional connectivity in remitted schizophrenia. Neuroscience Letters, 630, 199-202.
16. Sponheim, S.,, Clementz, B., Iacono, W., Beiser, M.(1994). Resting EEG in first-episode and chronic schizophrenia. Psychophysiology, 31(1), 37-43.
17. Tikka, S., Yadav, S., Nizamie, S., Das, B., Tikka, D., & Goyal, N. (2014). Schneiderian First Rank Symptoms and Gamma Oscillatory Activity in Neuroleptic Naïve First Episode Schizophrenia: A 192 Channel EEG Study. Psychiatry Investigation, 11(4), 467.
18. Venables, N., Bernat, E., & Sponheim, S. (2008). Genetic and Disorder-Specific Aspects of Resting State EEG Abnormalities in Schizophrenia. Schizophrenia Bulletin, 35(4), 826-839.