In the current study, we report that LFO amplitudes (i.e., ALFF and fALFF) were abnormal in patients with schizophrenia and schizoaffective disorder compared to healthy controls. fALFF measures were higher in controls than patients in the left cuneus and insula, right lingual gyrus and right caudate, and ALFF was higher in controls than patients in the right precuneus/cuneus and lingual gyrus, as well as right precentral gyrus. Patients had higher ALFF than controls in left hippocampus/parahippocampus, and higher fALFF than controls in the right parahippocampal and medial frontal gyri. Thus, our data suggest that there are abnormalities in LFOs in patients with schizophrenia. The current findings add to a literature suggesting abnormalities of neural synchrony in schizophrenia (Uhlhaas et al., 2008
) and extend these findings to the LFO domain. Whether these findings are related to the oscillatory hierarchies observed in higher frequency ranges (i.e., between delta, theta, and gamma ranges) remains an important question for future work.
Many of the areas in which reduced LFO amplitude in patients with schizophrenia was observed are sensory and motor regions, although deficits also were found in right precuneus, cerebellum, caudate, and paralimbic regions. These are consistent with deficits in low level visual and auditory sensory functions in schizophrenia (Javitt et al., 2000
; Rabinowicz et al., 2000
; Butler and Javitt, 2005
; Butler et al., 2005
), as well as with deficits in motor control seen in patients with schizophrenia (Bilder et al., 2000
; Bilder et al., 2002
). It may be speculated that the right striatal findings might relate to deficits in reward processing and/or stimulus saliency observed in schizophrenia (Murray et al., 2008
), and possibly to dulling of emotional expression in patients (Tremeau, 2006
), although this would have to be addressed in a study in which both behavioral measures and resting state measures are obtained. Cerebellar abnormalities have also been commonly observed in schizophrenia (Rusch et al., 2007
; Andreasen et al., 1999
The areas of increased amplitude in schizophrenia are in some of the same frontal and temporal regions previously associated with abnormal function in the disorder. The finding of increased fALFF in frontal regions is consistent with data showing that frontal regions may be dysregulated in the context of working memory task performance (Callicott et al., 2003
). Moreover, they may also be consistent with the notion that dysregulation of medial frontal regions is associated with self-directed thoughts, with the consequence that the source of internal and external stimuli could become confused, which may provide a neurophysiological basis for hallucinations (Whitfield-Gabrieli et al., 2009
). This would have to be verified in future work. These frontal increases were not detected by the ALFF measure. Taking into account the total power over the entire spectrum present in the BOLD signal, fALFF is the normalized ALFF, which can provide a more specific measure of low-frequency oscillatory phenomena.
We also observed hyperactivity in hippocampus/parahippocampus. These results are consistent with findings from other work showing hippocampal hyperactivity in patients with schizophrenia (Heckers, 2001
; Krieckhaus et al., 1992
; Lodge and Grace, 2007
). Some papers have suggested a left lateralization to this hyperactivity (Gur, 1978
). It is unclear why this abnormality was left sided for the ALFF measure and right sided for the fALFF measure. This difference will be the topic of future work.
The reduced LFO in the precuneus is consistent with abnormalities in the DMN in schizophrenia (see references in Introduction). The decrease there is in contrast with increased fALFF in medial prefrontal cortex, which is also a part of the DMN. It is possible that these differential effects on amplitude in patients are related to the disruption in coordination among elements of the DMN in schizophrenia (e.g., Whitfield-Gabrieli et al., 2009
; Garrity et al., 2007
Left caudate head fALFF was increased in patients whereas right caudate body fALFF was lower in patients than controls. This dissociation might be attributable to segregation of neural circuitry in the head vs. body of the caudate (Alexander et al., 1986
). Subsequent work has shown that activation in the head of the caudate is associated with feedback of performance and that activation in the body of the caudate is more involved in learning stimulus-category associations (Seger and Cincotta, 2005
; Cincotta and Seger, 2007
It is potentially interesting that the LFOs in the areas in which patients showed increases and decreases in ALFF and fALFF were negatively correlated in the patient group. This may suggest mutual inhibition among these sets of regions. Further work using diffusion tensor imaging might be particularly fruitful to examine this hypothesis.
Our findings for fALFF are similar to those in children with ADHD (Zang et al., 2007
) in showing increases in the right anterior cingulate cortex. It may be that some of the disinhibitory behaviors seen in both of these disorders might be shared by a common neural substrate. This raises interesting questions from a neurodevelopmental perspective, as well as from a nosological perspective that can be examined in future studies.
These current results were obtained under resting conditions. It would be important to determine whether similar results would be obtained under task conditions, or indeed whether differences in resting fALFF or ALFF would predict task induced signal changes. This issue can and should be examined in studies in which both resting and task-based fMRI are collected. It is possible, for example, that resting LFOs represent a neuronal activity baseline that can be entrained for the purposes of sensory selection of task-relevant stimuli (Schroeder and Lakatos, 2009
Several demographic variables were correlated with abnormalities in LFO. Illness duration was negatively correlated with ALFF in the patient deficit region. It is not known whether this might be a marker for possible progression of the illness. This could be addressed in longitudinal studies. In controls, age was negatively correlated with fALFF in the region in which patients showed heightened fALFF. It is unclear what the effects of age are in ALFF and fALFF. This should be addressed in future work.
The current study has several limitations. One of these is the unconstrained nature of the resting state. We should point out in this regard that Fransson (2006)
examined so-called task-unrelated thoughts (concerning inner speech, imagery, planning for the future, episodic memory, and task-unrelated attention) and found that these were correlated with neither activation during a working memory task nor with resting state activation in the default mode network. Moreover, we have shown short- and long-term test–retest reliability in the moderate to high range for resting state functional connectivity (Shehzad et al., 2009
). We also have shown that both fALFF and ALFF are reliable over time (Zuo et al., in press
). Secondly, patient participants all had chronic schizophrenia or schizoaffective disorders and were all on antipsychotic medication. However, neither fALFF nor ALFF differences were related to medication dosages. Subjects were instructed to keep eyes closed during the resting scan, but it is possible that there were group differences in the extent to which individual subjects might have failed to comply with this instruction. Future studies would benefit from the use of eye tracking or visual monitoring equipment during the resting state session. Finally, we examined only one frequency range. It would be important to examine other frequency ranges and their interrelationships, especially because these frequencies may exhibit a hierarchical structure (Lakatos et al., 2008
; Sirota et al., 2008
The current findings show that ALFF and fALFF are abnormal in schizophrenia, with multiple areas showing either increases or decreases in these measures. Although the relationship between these LFO measures and resting state functional connectivity (RSFC) remains to be determined, these group differences may form the basis for some of the RSFC deficits seen in schizophrenia (Garrity et al., 2007
; Bluhm et al., 2007
; Jafri et al., 2008
; Hoptman, et al., in press
). Moreover, our findings may have implications for abnormalities in the oscillatory hierarchy in schizophrenia that has been proposed by a number of different investigators, as noted above. The areas in which decreases were found are consistent with other data showing deficits in motor and low level sensory processing in this disorder, as well as with deficits in reward sensitivity in patients with schizophrenia. The areas in which increases were found are relevant to internally directed thought, which might lead to an inappropriate emphasis on internally generated stimuli, as well as confusion as to their source. For this reason, it has been postulated that this dysregulation might be associated with hallucinations. Further work will be needed to examine the functional significance of these findings.