A Review of Neuroimaging in Obstructive Sleep Apnea€¦ · Journal of Clinical Sleep Medicine, Vol. 2, No. 4, 2006 461 Obstructive sleep apnea (OSA) is a common clinical sleep dis-

Journal of Clinical Sleep Medicine, Vol. 2, No. 4, 2006 461

Obstructive sleep apnea (OSA) is a common clinical sleep dis-order that affects between 2% and 4% of middle-aged women

and men.1 Relatively increased prevalence rates have been report-ed in older adults2,3 and African Americans.4,5 OSA is associated with significant medical, cognitive, and psychological sequelae, including excessive daytime sleepiness, increased risk for motor vehicle accidents, depression, anxiety, obesity, insulin resistance, hypertension, and increased risk for vascular disease (e.g., see references 6-12). OSA is clinically characterized by chronically fragmented sleep and intermittent hypoxemia, defined as repeated episodes of deoxygenation that alternate with episodes of reoxy-genation. Sleep fragmentation may have detrimental effects on daytime functioning as a result of excessive daytime sleepiness.13 Lanfranchi and Somers14 have suggested that OSA-related hypox-emia is associated with an increase in sympathetic vasoconstric-tion and a coinciding decrease in vascular protective mechanisms, which in turn result in changes to the structure and function of the blood vessel. Aspects of this model have been incorporated

into theories of cognitive dysfunction in individuals with OSA.8,15 Although both sleep fragmentation and hypoxemia are believed to contribute to the adverse effects of OSA, their relative indepen-dent contribution and interaction remain poorly understood. Investigations applying neuroimaging methodologies have made important contributions toward improving our understand-ing of brain structure and function in individuals with OSA. Find-ings from neuroimaging studies may compliment more-tradi-tional sleep-assessment techniques, such as polysomnography, by providing unique information regarding brain structure, function, and metabolite composition. This in-vivo technique can clarify abnormalities in neural control of respiratory function in OSA by examining the relationship between respiratory challenges and brain function measured in the magnetic resonance imag-ing (MRI) scanner. Neuroimaging can also help identify neural abnormalities associated with vascular function in OSA. Finally, neuroimaging studies may aid the investigator in the identification of those individuals with OSA who are at greatest risk for poor outcome by examining relationships between brain integrity and functional response to treatment. These results may subsequent-ly serve as a potent clinical motivator for individuals with OSA who are struggling with treatment adherence. There are, however, limitations to the use of neuroimaging techniques. Some forms of neuroimaging, such as positron emission tomography, are in-vasive and not easily repeated over time. In addition, research-ers frequently utilize a wide range of imaging methodologies and post data-processing techniques that may make it difficult to compare results across studies. These results are often cross-sec-

A Review of Neuroimaging in Obstructive Sleep ApneaMolly E. Zimmerman, Ph.D.; Mark S. Aloia, Ph.D.

Department of Psychiatry and Human Behavior, Brown Medical School, Providence, RI

REVIEW ARTICLES

Study Objectives: The authors reviewed neuroimaging studies of obstructive sleep apnea (OSA) to summarize findings, evaluate their contribution to a current understanding of the neurophysiology of the disorder, and propose directions for future research.Method: Manuscripts were identified using the National Institutes of Health PubMed literature search system. Search terms included ob-structive sleep apnea, sleep apnea, imaging, neuroimaging, magnetic resonance imaging, MRI, functional magnetic resonance imaging, fMRI, magnetic resonance spectroscopy, and MRS. Inclusion criteria required that research articles (1) were written in English, (2) exam-ined an adult population, and (3) focused on imaging of the brain.Results: Support for structural abnormalities is mixed, but converg-ing evidence suggests the hippocampus may be atrophic in patients with OSA. Neurochemical evidence is supportive of white-matter im-pairment in OSA, particularly in the frontal lobes. Functional studies utilizing respiratory challenges report widespread neural differences in motor, sensory, and autonomic brain regions. Functional neuroimag-

ing cognitive challenges have reported either a lack of brain activation in dorsolateral prefrontal cortex or increased neural response in fron-tal lobe, cingulate, thalamus, cerebellum, and juncture of parietal and temporal lobes, depending on the cognitive task employed.Conclusions: The current literature examining neuroimaging-derived neural correlates in patients with OSA has made many important preliminary contributions. Future studies would be strengthened by consideration of potential moderating participant characteristics, such as sex, age, education, OSA severity, and comorbid conditions. Ad-ditional investigation employing neuroimaging techniques is needed to advance our understanding of the neurophysiology of OSA.Keywords: Obstructive sleep apnea, sleep-disordered breathing, neu-roimaging, magnetic resonance imaging, magnetic resonance spec-troscopy, functional magnetic resonance imaging, MRI, MRS, fMRICitation: Zimmerman ME; Aloia MS. A Review of Neuroimaging in Ob-structive Sleep Apnea. J Clin Sleep Med;2(4):461-471.

Disclosure StatementDrs. Zimmerman and Aloia have indicated no financial conflicts of interest.

Submitted for publication March 8, 2006Accepted for publication June 23, 2006Address correspondence to: Molly E. Zimmerman, PhD, Albert Einstein Col-lege of Medicine, Department of Neurology, 1165 Morris Park Ave. Bronx, NY 10461; Tel: (718)430-3919; Fax: (718)430-3870; Email: [email protected]


tional and correlative in nature, thereby prohibiting establishment of cause-and-effect relationships. Finally, the majority of modern neuroimaging techniques do not allow the continuous monitoring of brain function that would be ideal for an examination of sleep disorders. Despite these drawbacks, we believe that the applica-tion of neuroimaging techniques to the study of OSA represents an emerging field of inquiry that may help elucidate fundamental pathophysiologic mechanisms and clinical outcomes of the dis-order. The goal of the current review is to summarize findings of existing structural, neurochemical, and functional neuroimaging studies in individuals with OSA and to evaluate their contribu-tion to an improved understanding of the disorder. Limitations and considerations for future research are also discussed.

METHODS

Publications were identified using the National Institutes of Health National Library of Medicine PubMed literature search system. Additional references included within identified articles were also evaluated for inclusion in the review. Search terms in-cluded obstructive sleep apnea, sleep apnea, imaging, neuroim-aging, magnetic resonance imaging, MRI, functional magnetic resonance imaging, fMRI, magnetic resonance spectroscopy, and MRS. Different permutations of these search terms yielded be-tween 2 (obstructive sleep apnea and MRS) and 392 (sleep apnea and imaging) articles that were considered for review. Inclusion criteria required that original research articles (1) were written in English, (2) examined an adult population, and (3) focused on imaging of the brain. Articles were excluded if they focused on a population of interest other than OSA or sleep-disordered breathing; examined children, adolescents, or preclinical samples; performed neuroimaging of the upper airway; were primarily focused on physical abnormalities thought to cause OSA (e.g., craniometaphyseal dysplasia, congenital choanal atresia); or spe-cifically examined brain function following surgical treatment interventions (e.g., extended uvulopalatal flap surgery, mandibu-lar advancement). Published abstracts from national and interna-tional meetings and review articles were also excluded. In total, 17 studies were included in the current review.

REVIEW OF FINDINGS

The Table provides sample characteristics, summary of find-ings, and brief comments on all reviewed OSA neuroimaging studies. A detailed discussion of major findings by neuroimaging modality is presented below.

Structural Neuroimaging

Structural MRI (sMRI) is an important tool that promotes the examination of neuroanatomic volumetric and morphometric ab-normalities that may be involved in pathologic processes. sMRI also provides an important context in which to consider both neu-rochemical and neurofunctional findings. Many recent advances have been made in MRI image acquisition and postprocessing techniques that have contributed to improved sensitivity and spec-ificity of sMRI findings. A major limitation of MRI, however, is that measurements are acquired from a static image that may not correlate well with functional abnormalities or clinical variables. Nonetheless, sMRI studies facilitate an enhanced understanding of neuroanatomic substrates that inform both the development of

neuropathologic models and the interpretation of neurochemical and functional examinations of the brain. Several investigators have sought to characterize the structural neuroanatomy of individuals with OSA using sMRI. An early study by Davies and colleagues16 qualitatively examined white matter hyperintensities in deep white matter and periventricular regions in 45 patients with moderate to severe OSA and 45 con-trols who were closely matched on sex, body mass index (BMI), alcohol and cigarette use, hypertension, and history of heart dis-ease. Participants were also given ambulatory blood pressure monitors. Analyses revealed higher nighttime and daytime dia-stolic blood pressure and higher nighttime systolic blood pressure in patients with OSA, compared with controls. However, there were no group differences on any sMRI-derived qualitative indi-cator of subclinical cerebrovascular disease; indeed, both groups exhibited a high prevalence of deep white matter and periven-tricular hyperintensities. Given existing reports of associations among hypertension, stroke risk, and MRI abnormalities, the in-vestigators suggested that their unexpected finding indicated that causes of elevated blood pressure in patients with OSA may be differentially related to vascular risk compared to causes of el-evated blood pressure in healthy subjects. Macey and colleagues17 performed a quantitative sMRI analy-sis of gray and white matter in patients with OSA. They employed a voxel-based morphometric analytic technique in 21 men with OSA and 21 controls. Voxel-based analyses differ from manual volumetric methods in that they involve an automated compari-son of gray matter concentrations between 2 groups of interest on a voxel-by-voxel basis using MRI images that have been spatially normalized into stereotactic space.18 Although this technique al-lows the investigator to rapidly acquire a large amount of data across the entire brain, image-normalization processes may distort smaller regions of interest. The large number of resulting statisti-cal comparisons should also be controlled in data analyses. Using these techniques, the investigators reported regional gray matter loss in the frontal cortex, parietal cortex, temporal lobe, anterior cingulate, hippocampus, and cerebellum of patients with OSA. The extent of volumetric decline was found to be related to the se-verity of OSA, with patients with more severe OSA demonstrat-ing the greatest amounts of gray matter volume loss. Total gray matter volume decreased with age in controls but not in patients with OSA. This study was the first comprehensive, quantitative examination of MRI-derived neuromorphometry in OSA. These findings served to stimulate future neuromorphometric investiga-tions of OSA and provided an important initial framework for hypothesis generation. General interpretation of the results was limited, however, by the inclusion of patients with OSA with a wide range of comorbid conditions (e.g., medical and psychiatric disorders), as well as analyses of data that were uncorrected for multiple comparisons. A second comprehensive morphometric analysis of brain struc-tures in OSA was conducted by O’Donoghue and colleagues.19 This study examined 27 patients with untreated severe OSA and 24 age-matched controls using voxel-based morphometry and manual tracing of hippocampus, temporal lobe, and whole brain. Twenty-three patients with OSA were rescanned following 6 months of positive airway pressure (PAP) treatment, with aver-age adherence of 5.8 hours of use per night. Participants were excluded who had a history of cerebrovascular disease, diabetes, central nervous system disorder, alcohol or illicit drug use, or cur-

ME Zimmerman and MS Aloia


Table—Sample Characteristics and Major Findings of Reviewed OSA Neuroimaging Studies

First author, Imaging OSA Subjects Healthy Subjects Major Findings Commentsyear published Modality No. Age, y OSA severity No. Age, yDavies, 2001 sMRI 45 52(10) > 10 episodes of > 4% 45 52(10) No group differences in deep Men only fall in SaO2/night white matter or periventricular hyperintensitiesMacey, 2002 sMRI 21 49(11) AHI: 34(20) 21 47(11) OSA: ↓ gray matter in anterior Men only, included cingulate, hippocampus, patients with comorbid cerebellum, and frontal, parietal, conditions bid and temporal lobes O’Donoghue, 2005 sMRI 27 46(10) AHI: 72(17) 24 43(9) No group differences in gray 23 patients re-examined matter in any regions of interest; after 6 months PAP slight decrease in whole brain treatment volume in patients with OSA after 6 months of PAP treatmentMorrell, 2003 sMRI 7 50 (28-65) AHI: 28 (25-40) 7 50 (28-65) OSA: ↓ gray matter in left Men only hippocampusGale, 2004 sMRI 14 52(11) RDI: 84(18) 36% OSA group: ↓ hippocampus Comparison group: patients with CO poisoningDing, 2004 sMRI No WMD: No WMD: AHI:10 789 No differences in participants with SHHS 78(4); WMD: (12); WMD: and without brainstem white matter 79(5) AHI: 9 (12) disease in AHI, central or obstructive sleep apnea index, or % sleep time < 90% oxygen saturation; inverse relationship between frequency of arousals and white matter disease Robbins, 2005 sMRI 77(4) AHI: 11 843 Participants with brainstem white matter SHHS, disease progression more likely to longitudinal exhibit increase in central apneas dataKamba, 1997 MRS 23 49(13) AI: < 20 (mild), 15 46(18) Moderate to severe OSA: ↓ NAA/ Did not control > 20 (mod) Cho ratio in white matter compared for comorbid to mild OSA patients and controls disordersKamba, 2001 MRS 55 47(13) AHI: 44(30) Significant negative relationship No control between OSA severity and NAA/Cho group ratio in white matter Bartlett, 2004 MRS 8 47(16) RDI: 45 (43) 5 52(16) OSA: ↑ NAA/Cre, ↓ Cre in left Men only. hippocampus, associated with ↓ cognitive cognitive performance and OSA assessment severity Alchanatis, 2004 MRS 22 49(10) AHI: 71(19) 10 43(11) OSA: ↓ NAA/Cre and Cho/Cre Men only ratios, NAA, and Cho in frontal white matter Henderson, 2002 fMRI 8 44(4) RDI: 42 15 45(3) Group signal differences in Valsalva cerebellum, insular cortex, maneuver, hippocampus, cingulate, precentral men only gyrus, and frontal, temporal, and parietal cortices Harper, 2003 fMRI 10 46(12) AHI: 38(27) 16 47(10) Group signal differences in Cold pressor cerebellum, insular cortex, challenge, hippocampus, cingulate, frontal men only cortex, thalamus, and medulla/ midbrain Macey, 2003 fMRI 9 45(12) AHI: 40(28) 16 4(11) Group signal differences in Expiratory cerebellum, insular cortex, loading hippocampus, limbic regions, challenge, frontal cortex, thalamus, and men only midbrain/ponsMacey, 2006 fMRI 7 46 (5) AHI: 42(11) 11 47(3) Timing delay in basal ganglia in Inspiratory patients with OSA; group differences loading in cerebellum, insular cortex, challenge, men hippocampus, limbic regions, only. supplementary motor areas, temporal cortex, thalamus, and basal ganglia

Neuroimaging in OSA


rent use of psychoactive medications. Using an optimized post data-processing approach with adjustments for multiple compari-sons, the investigators found no evidence of gray matter change in any regions of interest in patients with OSA. Manual tracings were employed to reduce sources of error associated with auto-mated post data-processing techniques, such as masked effects due to normalization of brain tissue. Application of this technique similarly failed to reveal any structural differences between par-ticipants with OSA and controls. A slight decrease in whole brain volume, however, was observed in patients with OSA following 6 months of PAP treatment. In an additional attempt to replicate the earlier findings of Macey and colleagues,17 a posthoc analysis was conducted in which the data were reanalyzed using the same version of a postprocessing program employed by Macey et al, with consistent negative results. The authors suggested that the findings of Macey and colleagues may have resulted, in part, from inclusion of patients with comorbid medical and psychiatric con-ditions, analyses of data using a lenient threshold of signal detec-tion, and failure to correct for multiple comparisons. Additional published comments from both groups19,20 provide a provocative discussion of methodologic and statistical differences in neuro-imaging study design and data analysis. Macey and colleagues20 argue that their positive findings reflect the utilization of a statisti-cal threshold that allowed detection of known age-related effects on the brain, whereas O’Donoghue and colleagues19 state that their negative findings result from the utilization of an MRI scan-ner with a stronger field strength, a homogeneous sample, and statistical correction for both multiple comparisons and age. Both groups appear to agree that the effects of OSA on brain structure are relatively small and may not be evident when highly stringent and conservative statistical methods are applied. These compel-ling studies and their respective follow-up arguments serve to highlight the variability that may exist in study cohorts and post-processing techniques. Further, they illustrate an urgent need for additional well-controlled morphometric studies of patients with

a wide range of OSA severity to replicate findings and further investigate volumetry associated with OSA. Two additional studies have also examined structural brain in-tegrity in OSA. Morrell and colleagues21 conducted a cross-sec-tional voxel-based morphometric study on 7 patients with newly diagnosed moderate OSA and 7 controls matched for handedness and age. Results indicated that patients with OSA had significantly smaller left gray matter hippocampal volumes, as compared with controls. There was no difference between the groups in any other gray matter brain region or in total gray matter volume. Cognitive correlates of hippocampal volume were examined in a study by Gale and Hopkins.22 To examine the effects of hypoxemia on the brain, MRI and comprehensive neuropsychological assessment were performed on patients with severe OSA and patients with carbon monoxide poisoning. Time between carbon monoxide ex-posure and MRI scan or neuropsychological testing was 22.4 + 13.8 months for the patients with carbon monoxide poisoning, whereas the patients with OSA underwent testing within several days of diagnosis and before treatment with PAP. There were no racial or ethnic differences between the 2 patient groups. Results revealed that 36% of patients with OSA exhibited hippocampal atrophy in the context of normal ventricle-to-brain ratios. Brain atrophy was identified through comparisons with a previously es-tablished normative sample of quantitative brain volumes.23 The presence of hippocampal atrophy in the OSA group was signifi-cantly associated with cognitive measures of nonverbal memory and information processing. There was also a statistically signifi-cant relationship between baseline oxygen saturation and hip-pocampal volume in patients with OSA. Interestingly, the OSA group generally performed relatively better than the carbon mon-oxide poisoning group across different neuropsychological tests, although they did exhibit impairments on selective tests, particu-larly executive function, when their performance was compared with that of a normative sample. White matter integrity has also been a focus of sMRI popula-

Table—Continued

First author, Imaging OSA Subjects Healthy Subjects Major Findings Commentsyear published Modality No. Age, y OSA severity No. Age, yThomas, 2005 fMRI 16 40(7) RDI: 58(16) 16 38(6) OSA: absence of activity in Working memory dorsolateral prefrontal cortex cognitive challenge, both before and after treatment; patients re-examined patients with OSA performed after 8 weeks PAP more poorly on cognitive task, treatment compared with controls Ayalon, in press fMRI 12 44(12) AHI: 35(21) 12 43(9) Intact verbal learning performance Verbal list in patients with OSA associated learning cognitive with increased brain activity in right challenge, 11 inferior frontal gyrus, middle frontal men, 1 woman gyrus, cingulate gyrus, junction of the in each group inferior parietal and superior temporal lobes, thalamus, and cerebellum

Age is presented as the mean (SD) for all studies, with the exception of Morrell et al, which is presented as the mean (range), and Bartlett et al, which is presented as the median (interquartile range).Obstructive sleep apnea (OSA) severity measures are presented as the mean (SD) for all studies, with the exception of Bartlett, which is presented as the median (interquartile range).sMRI refers to structural magnetic resonance imaging; MRS, magnetic resonance spectroscopy; fMRI, functional magnetic resonance imaging; PAP, positive airway pressure; AHI, apnea hypopnea index; RDI, respiratory disturbance index; AI, apnea index; NAA, N-acetylaspartate; Cho, choline; Cre, creatine; CO, carbon monoxide; SHHS, the Sleep Heart Health Study, an epidemiologic population-based sample of older adults > age 68 years.



tion-based studies. The relationship between white matter disease and sleep-disordered breathing was examined in an epidemiologic dataset comprised of 789 older community-dwelling adults over the age of 68 years from the Sleep Heart Health Study.24 The Sleep Heart Health Study is a population-based study designed to ex-amine the relationship between cardiovascular disease and sleep-disordered breathing. Evaluation of white matter disease was conducted in the midbrain, pons, and medulla. Sleep-disordered breathing was defined using an apnea-hypopnea index, arousal index, and both obstructive and central apnea indexes. Polysom-nography was obtained from 1995 to 1998, and MRI was obtained from 1997 to 1998. Blood pressure, weight, BMI, smoking history, age, alcohol use, history of coronary heart disease, and diabetes history were examined as possible covariates. The results indicat-ed that approximately 27% of participants in the sample exhibited evidence of white matter disease in the brainstem, predominantly in the pons. There were no differences between participants with and without white matter disease in apnea-hypopnea index, cen-tral or obstructive sleep apnea indexes, or percentage of sleep time with an oxygen saturation less than 90%. However, individu-als with white matter disease did exhibit fewer arousals per hour of sleep; these variables maintained a significant association even after adjusting for age, sex, race, BMI, and blood pressure. The authors concluded that their negative findings may be related to the relatively mild severity of sleep-disordered breathing in their population-based sample, survival biases, or the high comorbid-ity of other vascular risk factors that may be present in a sample of older adults. Regarding the unexpected inverse relationship observed between frequency of arousals and prevalence of white matter disease, the authors speculated that the arousal response may be protective against white matter disease in the brainstem. However, they also cautioned that the clinical significance of this association should be more fully investigated. Longitudinal data from the Sleep Heart Health Study were uti-lized to examine temporal relationships between sleep-disordered breathing and MRI-derived indexes of white matter infarcts in the brain.25 The sample was comprised of 843 individuals with a mean age of 77 years who received MRI scans in 1992 to 1993 and 1998 to 1999 and polysomnography in 1995 to 1997. Exclusion criteria for the Sleep Heart Health Study were that participants could not be in active PAP treatment, have a tracheotomy, or be on cur-rent home oxygen therapy. Characterization of sleep-disordered breathing included consideration of both central and obstructive apneas. The results indicated that participants who displayed pro-gression in white matter disease on MRI were significantly more likely to have an increase of central, but not obstructive, apneas compared with individuals without evidence of progression of white matter infarcts. The authors concluded that this relationship suggests that central sleep apnea either contributes to the progres-sion of white matter disease or may be a marker of subclinical cerebrovascular disease.

Magnetic Resonance Spectroscopy

Magnetic resonance spectroscopy (MRS) is an imaging tech-nique that permits investigation of neuronal cellular chemical activity through examination of neurotransmitters and amino ac-ids. In existing MRS studies of populations with OSA, spectral resolutions of N-acetylaspartate (NAA), choline(Cho), creatine (Cre), and myo-inositol (mI) metabolites have been examined.

NAA, an amino-acid derivative, is primarily located in neurons and is thought to be a marker of neuronal viability.26,27 Reduc-tions of NAA may reflect neurodegeneration.27 Abnormal levels of Cho suggest inflammation, cellularity, and membrane degrada-tion associated with demyelination.26,28 A decrease in the ratio of NAA to Cho has been utilized as an indicator of cerebral meta-bolic injury, such as gliosis and impairment of neuronal and axo-nal function.29-31 Creatine reflects biochemical energy reserves of glia and neurons and is presumed to be stable in the context of pathologic processes.26 As a result of its relative stability, creatine is frequently utilized as a control in metabolic ratio values (e.g., Cho/Cre). Finally, an increase in mI metabolites may be an in-dicator of glial-cell proliferation or gliosis.26,28 Because spectral data from MRS studies are acquired from relatively small levels of neural metabolites, regions of interest are often limited in size. Nonetheless, MRS is a useful neuroimaging tool for the study of OSA because it provides a measure of cerebral metabolic change that may reflect pathologic insults to brain integrity. Kamba and colleagues conducted 2 MRS studies in patients with OSA. A 1997 brief report32 applied MRS in patients with mild OSA, moderate to severe OSA, and controls to obtain NAA/Cho, NAA/Cre, and Cho/Cre ratios in both cerebral cortex and white matter. Many patients with OSA had comorbid medical conditions, including hypertension, cardiac hypertrophy, bron-chial asthma, nasal allergy, tonsillectomy, cervical spondylosis, and diabetes mellitus. The findings revealed a significant group difference in the NAA/Cho ratio for cerebral white matter, with a significantly lower ratio in the patients with moderate to severe OSA, compared with those with mild OSA and controls. The au-thors concluded that metabolic changes occur in normal-appear-ing white matter of patients with moderate to severe OSA. These results represent an important early contribution to the study of OSA using neuroimaging techniques that provides support for possible cerebral damage associated with the disorder. However, the findings are limited in their specificity because the authors did not control for common comorbid medical conditions in OSA that may be independently related to cerebral ischemia and metabolic impairment. A follow-up MRS study33 was implemented to address the effects of potentially confounding comorbid conditions on the examination of cerebral metabolic activity in OSA. Fifty-five patients with severe OSA underwent MRS to acquire NAA/Cho ratios for cerebral cortex and periventricular white matter. Patients were evaluated for the presence of hypertension, cardiac disease, diabetes mellitus, and hyperlipidemia. Patients with evidence of brain infarction, hemorrhage, or white-matter hyperintensity le-sions on MRI were excluded from analyses. Results indicated that the NAA/Cho ratio for cerebral cortex decreased with age. Al-though no significant relationship was found between severity of OSA and the NAA/Cho ratio for gray matter of the cerebral cor-tex, severity of OSA was found to have a significant negative as-sociation with NAA/Cho ratio for normal-appearing white matter that was independent of age and the presence of cardiac disease. The investigators reported that severity of OSA may be associ-ated with the degree of white-matter metabolic impairment and that this impairment may be related, in part, to comorbid cerebro-vascular risk factors. Although this investigation did not include a healthy control comparison group, the findings are generally supportive of those previously obtained by the authors. Further-more, they extend those findings by emphasizing the importance

Neuroimaging in OSA


of consideration of comorbid medical conditions that may also have an effect on metabolic findings as measured by MRS. Bartlett and colleagues34 conducted a preliminary spectroscopy study of 8 men with severe OSA and 5 age-matched controls. Data were obtained for NAA/Cre and Cre compounds in the left hippocampus. Participants also underwent cognitive assessment of both vigilance and processing speed prior to and following the spectroscopy scan. The findings revealed a significant increase in NAA/Cre ratio in the left hippocampus of patients with OSA that the authors reported was likely the result of an observed decrease in the Cre metabolite. In the OSA group, NAA/Cre was signifi-cantly correlated with arousal index but not with total respiratory disturbance index or average oxygen desaturation. Creatine was significantly associated with decreased vigilance performance pri-or to and following the MRS scan, whereas NAA was associated with vigilance performance only prior to the scan. The authors did not specifically examine differences in cognitive performance be-tween the patient and control groups. They concluded from their findings that decreased Cre levels in the hippocampus of patients with OSA may be related to intermittent hypoxemia associated with the disorder because creatine is involved in energy homeo-stasis and has been shown to improve performance on cognitive tests.35 Although obtained from a small sample size, these results offer additional evidence of both metabolic abnormalities in OSA and hippocampal dysfunction. A novel contribution of this study includes identification of a significant association of metabolic levels and cognitive test performance, supporting the continued investigation of these relationships in patients with OSA. A recent MRS study36 examined brain metabolism in the pre-frontal cortex, parieto-occipital, and frontal periventricular white matter of patients with untreated severe OSA and age- and sex-matched healthy controls. MRS was utilized to obtain spectral data for NAA/Cho, NAA/Cre, Cho/Cre, mI/Cre ratios, and abso-lute concentrations of NAA, Cho, Cre, and mI metabolites. The results revealed a decrease in NAA/Cre and Cho/Cre ratios and absolute concentrations of NAA and Cho in the prefrontal white matter of patients with OSA, compared with controls. Although 8 of the 22 patients reported a history of cerebrovascular risk fac-tors, analyses of a subgroup of patients without comorbid medical conditions revealed that the findings were essentially unchanged. Taken together, these MRS-derived findings provide important additional support for the presence of neurochemical abnormali-ties thought to reflect white matter impairment in individuals with moderate to severe OSA.

Functional Neuroimaging

Although there are several different functional neuroimaging approaches, to our knowledge, existing functional neuroimaging studies of OSA have utilized only functional MRI (fMRI). fMRI allows the researcher to examine cerebral activation in response to an external probe. fMRI images are generally acquired using the blood-oxygen level-dependent (BOLD) technique. This non-invasive, high spatial and temporal resolution technique enables acquisition of images that are dependent on the sensitivity of the MR signal to excesses of cerebral blood flow associated with an increase of synaptic activity in the brain. Functional neuroimag-ing techniques are ideal for investigations of acute changes in the brain associated with performance of various challenges. Coupled with this, however, is the limitation that fMRI data can only re-

flect relative changes in neural activity that are specifically asso-ciated with those challenges.37 Furthermore, fMRI is an indirect measure of vascular function that may be compromised in indi-viduals with vascular pathology. These are crucial considerations when interpreting fMRI findings in both healthy and clinical populations. Current fMRI studies of OSA employ either respira-tory or cognitive challenges, described in greater detail below, designed to activate specific brain regions during image acqui-sition. In brief, respiratory challenges employ forced alterations of breathing effort to evoke changes in the autonomic nervous system similar to those experienced by individuals with breathing disorders, whereas cognitive challenges elucidate neural mecha-nisms associated with performance on complex mental tasks.

RESPIRATORY CHALLENGES

A complementary series of 4 fMRI studies investigated neu-ral function associated with respiratory challenge in patients with OSA. The first of these38 utilized the Valsalva maneuver, an ex-piratory challenge in which participants are instructed to exhale vigorously against a resistant force that causes a closed glottis. Clinically, the Valsalva maneuver has been used to assess auto-nomic nervous system function and treat cardiovascular disorders, such as angina pectoris and paroxysmal tachycardia.39 Prolonged expiratory effort results in increased upper airway pressure and an initial increase in blood pressure associated with expiratory strain. The maintenance of this strain results in a rapid decline in blood pressure, followed by an eventual drastic overcompen-sation.40 Changes in cerebral hemodynamics associated with the challenge include altered flow velocity in the middle cerebral ar-tery.39 Henderson and colleagues38 performed a series of 3 repeat-ed 18-second Valsalva maneuvers at a load pressure of 30 mm Hg during fMRI in patients with OSA and in controls. None of the patients with OSA was taking medications, and both patients and controls had similar BMIs. Absence of sleep-disordered breathing was verified in healthy controls through respiratory monitoring and electroencephalography conducted during sleep in the MR scanner. Analyses indicated that the Valsalva maneuver was as-sociated with group differences in regional neural signal inten-sity in the left inferior parietal lobe, left precentral gyrus, anterior superior temporal gyrus, superior frontal gyrus, posterior insular gyrus, cerebellum, anterior cingulate, and hippocampus. Group differences in neural response-time patterns were also noted in the cerebellum, posterior insula, anterior cingulate, and precentral gyrus. The authors concluded that the OSA-related regional spec-ificity of these results supported those of previously reported17 areas of gray matter volume loss. They also stated that these find-ings suggested the presence of neural functional reorganization that may mediate cardiovascular and respiratory mechanisms in patients with OSA performing the Valsalva maneuver. A second fMRI autonomic respiratory challenge41 was con-ducted in participants with OSA and controls using a cold pres-sor technique. This assessment of the integrity of the autonomic nervous system involves the application of a cold stimulus to a participant’s forehead to elicit an involuntary autonomic reflex that physiologically results in decreased breathing and heart rate and increased blood pressure. In the current study, the cold stimulus consisted of a 4oC bag of deuterium oxide. Participants in this sample overlapped with those of the previously described Henderson et al study.38 Several patients with OSA were unable



to participate in the study due to scanner weight limitations or claustrophobia. Cluster analyses indicated that patients with OSA exhibited decreased neural responses relative to controls in the ventral thalamus, ventral anterior insula, and hippocampus. Rela-tive signal increases in patients with OSA were observed in cere-bellum, anterior insula, medial frontal cortex, and cingulate gyrus. It was concluded that the cold pressor challenge produces OSA-related neural differences in multiple brain regions, particularly cerebellum and limbic areas. Macey and colleagues42 conducted an expiratory-loading breathing challenge to examine associated neural response in OSA. Expiratory-loading challenges require a sustained breathing effort that produces low positive airway pressure with few exag-gerated muscle exertions. In this study, the 90-second period of sustained expiratory-loading airflow differed from the repetitive, short, breathe-pause efforts associated with the Valsalva maneu-ver. Participants were instructed to sustain expiratory effort at a pressure of 10 mm Hg or greater, and all participants practiced the exhalation at least 20 minutes prior to initiation of the challenge. The participant sample overlapped with that of the 2 previously described studies.38,41 In the current sample, 3 patients with OSA were prescribed PAP treatment but did not use it the night before the scan. Analyses revealed group differences in neural signal in-tensity in several cortical and subcortical regions. A relative sig-nal decrease was observed in the subjects with OSA in the right insula, left anterior cingulate, and middle frontal gyrus, whereas a relative signal increase was observed in the right hippocampus, ventral midbrain, left dorsal midbrain, and right ventral pons. Variability in neural response in the patients with OSA was noted in the amygdala, cerebellum, and posterior insula. The authors suggested that these brain regions may mediate neural response to resistive breathing challenges and may play a role in the abnormal physiologic mechanisms associated with OSA. The most recent study43 in this series of fMRI respiratory chal-lenges examined inspiratory loading in 7 patients with OSA and 11 controls. Inspiratory challenges involve increased negative pressure on the upper airway resulting from prolonged periods of inspiration followed by brief expiration. This produces autonomic changes to blood pressure associated with alterations of breathing effort similar to those observed in individuals with OSA. In the current study, each inspiratory challenge consisted of 60 seconds of unrestricted breathing and 90 seconds of inspiratory loading at a pressure ranging between 6 and 15 mm Hg. The authors hy-pothesized that inspiratory loading would elicit different neural responses in patients with OSA and controls in motor regions and sensory areas that mediate autonomic stimulation. Similar to the results from the 3 other studies in this series, altered neural signal intensities in response to inspiratory loading were observed in pa-tients with OSA in multiple brain regions. The largest clusters of group differences were found in the basal ganglia and left insula. Signal differences were also observed in the medial cingulate cor-tex, right ventral posterior thalamus, anterior thalamus, right hip-pocampus, left medial temporal cortex, medial midbrain, and cer-ebellum. Response-timing alterations were exhibited in the basal ganglia. The authors concluded that timing and signal-intensity differences in these motor, sensory, and autonomic integration re-gions of the brain may underlie nocturnal pathologic breathing abnormalities in patients with OSA. It was also noted that many of the functional abnormalities observed in these studies were in brain regions that were shown previously to have volumetric gray

matter abnormalities.17 Considered in concert, the results from this series of respira-tory-challenge studies illustrate the importance of examination of neural activity in response to various breathing paradigms in pa-tients with OSA. Whether the observed brain abnormalities are a cause or effect of disordered breathing patterns, however, remains an issue for further study. Furthermore, the effect of treatment on neural function of the patient with OSA in response to a respi-ratory challenge also requires additional clarification. Although no patient with OSA used PAP the night before MR scanning, some patients were prescribed PAP treatment prior to enrollment in the study. Overall, the findings from these studies provide a vital contribution to an improved understanding of the effects of disordered breathing on brain function in the patient with OSA.

COGNITIVE CHALLENGES

Thomas and colleagues44 utilized a cognitive activation para-digm to examine brain activity in 16 patients with untreated OSA and 16 controls. Six patients with OSA were also rescanned fol-lowing a minimum of 8 weeks of compliant use of PAP treatment (> 7 hours/night or 100% of total sleep time). Patients reported excessive daytime sleepiness and a duration of OSA symptoms of at least 5 years. All participants were reported to be medically and psychiatrically healthy. The cognitive challenge utilized in this study was the 2-back verbal working memory task, which requires mental comparison of rapidly presented verbal stimuli to stimuli presented in 2 previous trials. Cognitive comparisons re-quire the participant to maintain a dynamic mental representation of previously presented stimuli in short-term working memory. Using a blocked data-acquisition design, the investigators report-ed an absence of dorsolateral prefrontal activation and decreased working-memory speed in patients with OSA, as compared with controls. An fMRI blocked design involves repeated exposure to 1 condition (A) alternating with repeated exposure to a sec-ond condition (B), resulting in an “AB block” design that can be presented sequentially during scan acquisition. Hypoxemia did not contribute to the finding in the patient group. Following PAP treatment, patients demonstrated an improvement in subjective sleepiness and partial recovery of posterior parietal activation. However, they also continued to exhibit an absence of dorsolat-eral prefrontal activation. The authors concluded that their data supported a functional anatomic model of excessive sleepiness; that is, altered functional neurocircuitry was demonstrated in the dorsolateral prefrontal cortex in patients with OSA, similar to neural activation patterns observed under conditions of sleep deprivation. Ayalon and colleagues45 recently conducted an fMRI study to examine brain activity associated with a cognitive challenge (verbal list learning) in patients with untreated OSA and controls. There were no differences between the patients and controls on BMI, blood pressure, age, education, or English fluency. The au-thors reported that the patients with OSA and controls performed similarly on the verbal list learning task across multiple indexes, including immediate and delayed recall and recognition memory. However, a significant task-associated increase in brain activa-tion in the patients with OSA was observed in the right inferior frontal gyrus, middle frontal gyrus, cingulate gyrus, junction of the inferior parietal and superior temporal lobes, thalamus, and cerebellum. Better verbal learning performance in patients with

Neuroimaging in OSA


OSA was associated with increased brain activation in the left in-ferior frontal gyrus and left supramarginal area, whereas poorer performance was associated with increased activation response in left inferior parietal lobe. The authors concluded that the unique pattern of cerebral activation in patients with OSA in the con-text of intact cognitive performance may reflect recruitment of additional neural resources consistent with an adaptive compen-satory response, similar to that seen in healthy adults following total sleep deprivation46-48 and elderly adults.49,50 Inconsistencies between these findings and those of Thomas and colleagues44 are likely due to differences in the cognitive-challenge paradigm and severity of OSA. The Thomas study utilized a working-memory cognitive task, whereas Ayalon and colleagues employed a test of verbal learning. Given that these 2 tasks presumably involve different cognitive processes, it is not surprising that each is as-sociated with varying patterns of neural signal abnormalities in the patient with OSA. Taken together, the findings from these functional neuroimaging cognitive-challenge studies provide an important framework for the development of additional studies that examine neural activity of patients with OSA.

CONCLUSIONS

Summary of Findings

The Table summarizes characteristics of the OSA sample and neuroimaging findings. Structural neuroimaging studies have provided inconsistent evidence for generalized changes in gray matter, which may be related to differences in individual study samples. More consistent findings, however, have been reported regarding decreased hippocampal volume. This abnormality has been hypothesized to be associated with the effects of chronic, intermittent, nocturnal hypoxemia, although this model has not been explicitly tested. Epidemiologic studies of structural white-matter integrity generally have not found evidence of abnormali-ties in community-dwelling older adults with undiagnosed breath-ing difficulties. OSA-associated dysfunction of the hippocampus is also sup-ported from the findings from at least one MRS study.34 Perhaps the most consistent finding of MRS studies, however, involves white-matter dysfunction, particularly in the frontal lobes. The lack of convergence between MRS and structural MRI white-matter findings should not be a primary concern, given vari-ability in methodologic approaches and sample characteristics. Specifically, negative structural white-matter studies were often composed of older adults from the general population, whereas spectroscopy samples contained middle-aged adults with clini-cally diagnosed OSA. In general, these results suggest that white-matter abnormalities may be present in selected individuals who are at the greatest risk of developing vascular disease as a result of chronic apneic events. Existing models of central nervous system dysfunction have suggested that vascular compromise and endo-thelial dysfunction in OSA may preferentially damage small ves-sels in the brain,14 which could result in small-vessel, white-mat-ter ischemia. Utilization of emerging MRI techniques that have been specifically developed to examine white-matter integrity in normal-appearing white matter, such as diffusion-weighted tensor imaging, may further advance our understanding of white-matter dysfunction in OSA. Functional neuroimaging studies have also made vital contri-butions to the OSA literature. fMRI respiratory challenges pro-

vide support for OSA-related differences in neural function in multiple brain regions involved in respiratory and cardiovascular control, including regions in the motor, sensory, and autonomic integration brain areas. Although respiratory challenges con-ducted during wakefulness do not directly mimic the physiology of sleep-disordered breathing, these studies nonetheless provide important insights into the neural mechanisms that may underlie the etiology or be involved in the physiologic consequences as-sociated with OSA. Results from the few fMRI studies utilizing cognitive probes in individuals with OSA are mixed. Task-related results dependent upon the activation paradigm were evident in all studies. Some findings were consistent with fMRI studies of sleep deprivation, demonstrating compensatory recruitment of brain regions in untreated individuals. Treatment-related findings require further replication, given that only 1 study to date has examined the effects of compliant PAP treatment in 6 individuals with OSA. Considered together, these neuroimaging findings provide support for the presence of OSA-associated neurofunctional and white-matter impairments, particularly in the frontal lobes and hippocampus. Such impairment is consistent with proposed mod-els of central nervous system and cognitive dysfunction in OSA implicating small vessel disease8 and prefrontal cortex.15 Given the relative paucity of experimental inquiry, however, additional studies are needed to more fully substantiate these models.

Limitations and Considerations

Several important limitations and considerations emerge from a review of the existing neuroimaging literature in OSA. Most no-tably, the vast majority of studies have included only male partici-pants, thereby limiting generalizability of findings to men with OSA. Although the clinical presentation of OSA is more com-mon in men than in women,51,52 the sex ratio of published studies does not generally reflect this prevalence. Consideration of the potential moderating effects of IQ and education is also impor-tant, particularly with respect to treatment-adherence studies and performance on cognitive challenges. Although prior studies have shown relationships between IQ and normal brain development53 and potentially protective effects of high IQ against OSA-associ-ated cognitive decline,54 only 2 studies in the current review22,45 reported measures of intellectual or educational achievement. Age may also moderate the effect of OSA on brain structure and func-tion. For instance, advanced age may mask the effects of OSA on cerebral integrity. Associations between sleep parameters and neuroimaging may be less robust among older adults, given the myriad of age-related comorbid factors that could affect this rela-tionship. Likewise, length of illness is a potentially confounding variable that is difficult to assess in the patient with OSA and, therefore, is difficult to adequately investigate. Several investiga-tors have specifically focused on treatment-naïve patients in an attempt to address this issue (e.g., see references 19, 21, 36, and 45), but more-direct assessment methods (e.g., bed-partner ques-tionnaire) may be necessary. Taken together, these considerations indicate that researchers should seek to adequately match OSA patients and healthy subjects on a wide range of demographic variables while considering individual differences in order to more clearly elucidate the pathophysiology of OSA. Contradictory results among several of the reviewed studies may be partially due to a failure to adequately examine the ef-



fect of common comorbid medical conditions that may also be associated with alterations in brain structure and function. For instance, hypertension is commonly reported in patients with OSA6 and has known independent effects on brain structure55-57 and function.58 fMRI studies that seek to examine hypertension as a potential mediator of the relationship between OSA and brain dysfunction should also consider the effect of hypertension on the BOLD signal, given its reliance on blood flow.59,60 Similarly, OSA severity and its relationship with other clinical symptoms is an important factor to consider when designing neuroimaging stud-ies. Several of the studies included in the current review included participants from the general population with subclinical symp-toms of OSA,24,25 rather than clinically referred OSA patients with more severe symptomatology. Although examination of patients with a wide range of OSA severity is necessary to gain a thorough understanding of the clinical and physiologic characteristics of the disorder, caution should be taken when comparing and inter-preting results across studies. There may be a critical threshold associated with the clinical expression of OSA-related pathologic insult to the brain. Similarly, patients with very severe OSA may be more likely to have comorbid illness that may independently compromise the brain. In a recent study of hypertension in the Wisconsin Sleep Cohort, investigators suggested that the rela-tionship between OSA and hypertension varied across levels of OSA severity.61 Clearly, continued examination of the complex relationship between disease severity and comorbid medical con-ditions is critical. Perhaps one of the most significant problems associated with neuroimaging in OSA involves the sampling bias that exists when conducting MRI scans on obese patients with OSA. Although BMI is often reported in the reviewed studies, the effect of BMI on observed findings is generally not explored or discussed. In a practical sense, many patients with OSA are unable to appropri-ately fit into the MRI scanner. It is therefore likely that OSA sam-ples included in neuroimaging studies represent only a subgroup of patients with the disorder. This limits interpretation regarding the interaction of OSA and obesity and/or any pathophysiologic differences in OSA among obese and nonobese patients. There will likely be methods to remedy this problem as continued efforts increasingly focus on the utilization of neuroimaging techniques in obesity, but investigators should be aware of this sampling bias until these issues are resolved. An additional limitation and consideration that emerges from a review of the OSA imaging studies is that, with the exception of Robbins et al’s study,25 all reviewed studies examined cross-sectional rather than longitudinal, data. Although cross-sectional studies provide valuable information for the investigation and characterization of a sample, longitudinal data may allow the researcher to more specifically examine the development of un-derlying neuropathologic processes associated with OSA. When designing a longitudinal study, the investigator may benefit from consideration of the feasibility of acquiring and comparing re-peated scans over time in the same participant using the same MR scanner. The effects of treatment on the brain of the OSA patient are also only beginning to be understood. Two studies in-cluded in the current review19,44 that examined change associated with treatment provide an excellent model for the consideration of PAP-adherence data when interpreting treatment outcomes. However, these studies included participants with high levels of nightly PAP adherence. Given the difficulties that many patients

experience maintaining recommended nightly PAP use,62-64 future neuroimaging studies investigating the effect of varying amounts of adherence on brain structure and function would be an impor-tant contribution to the OSA literature.

Future Research Directions

Although existing neuroimaging studies have made critical contributions to a scientific understanding of the underlying bio-logic mechanisms of OSA, several directions for future research emerge from a review of the literature:1. Explore individual differences and potentially moderating

effects of demographic variables on the neural structure and function of the patient with OSA, including sex, education, IQ, and BMI.

2. Describe the course of OSA and the effect of varying levels of treatment using both cross-sectional and longitudinal ex-perimental neuroimaging designs.

3. Examine the relationship between neuroimaging findings and more comprehensive assessments of cognitive and be-havioral outcomes in patients with OSA.

4. Expand the characterization of neurochemical abnormalities in the patient with OSA using MRS and examination of ad-ditional metabolites, such as glutamate.

5. Replicate fMRI studies using cognitive and respiratory chal-lenges in patients with OSA.

In summary, neuroimaging studies have broad scientific appeal for the study of OSA. We believe that it is becoming increasingly important to apply neuroimaging techniques to neuropathologic models of OSA with a priori working hypotheses. In this way, neuroimaging studies of OSA may serve to help clarify underly-ing biologic mechanisms associated with the onset and mainte-nance of the disorder. They may also help characterize response to treatment, at both therapeutic and subtherapeutic levels. Ob-jective findings from neuroimaging investigations may provide useful clinical motivators for patients with OSA struggling with treatment adherence. Whole-brain neuroimaging approaches represent a useful preliminary approach and should continue as attempts are undertaken to develop consensus on pathologic re-gions of interest in OSA. Theory- and model-based approaches, however, guide researchers toward rigorous scientific examina-tion of specific questions and encourage modification of existing theoretic models of OSA. This will aid in the development of complementary research ideas that promote innovative transla-tional research efforts. Convergence of multimodal neuroimaging efforts may also serve to inform theory development and refine-ment. Large-scale community studies may consider including sleep-related measures in their assessments to ensure that their participants exhibit normal sleep functions. Ongoing advances in imaging techniques will likely remedy some of the current limita-tions inherent in neuroimaging investigations of OSA. For ex-ample, magnetoencephalography, a noninvasive direct measure of magnetic fields generated by neural electrical activity, can more readily accommodate the obese patient. Similarly, optical imaging, although limited in spatial resolution, provides a mobile imaging technique that affords continuous imaging throughout sleep. Finally, consideration should be given to the development and continual review of standards for neuroimaging research in OSA. This will facilitate comparison of findings and help elimi-nate methodologic differences across studies.

Neuroimaging in OSA


ACKNOWLEDGEMENTS

The work presented in this paper was supported by NIH grant R01 HL 075366.

REFERENCES

1. Young T, Peppard PE, Gottlieb DJ. Epidemiology of obstructive sleep apnea: a population health perspective. Am J Respir Crit Care Med 2002;165:1217-39.

2. Ancoli-Israel S, Kripke DF, Klauber MR, Mason WJ, Fell R, Ka-plan O. Sleep-disordered breathing in community-dwelling elderly. Sleep 1991;14:486-95.

3. Duran J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apnea-hy-popnea and related clinical features in a population-based sample of subjects aged 30 to 70 yr. Am J Respir Crit Care Med 2001;163:685-9.

4. Redline S, Tishler PV, Hans MG, Tosteson TD, Strohl KP, Spry K. Racial differences in sleep-disordered breathing in African-Ameri-cans and Caucasians. Am J Respir Crit Care Med 1997;155:186-92.

5. Kripke DF, Ancoli-Israel S, Klauber MR, Wingard DL, Mason WJ, Mullaney DJ. Prevalence of sleep-disordered breathing in ages 40-64 years: a population-based survey. Sleep 1997;20:65-76.

6. Hedner J, Bengtsson-Bostrom K, Peker Y, Grote L, Rastam L, Lindblad U. Hypertension prevalence in obstructive sleep ap-noea and sex: a population-based case-control study. Eur Respir J 2006;27:564-70.

7. Aloia MS, Arnedt JT, Smith L, Skrekas J, Stanchina M, Millman RP. Examining the construct of depression in obstructive sleep apnea syndrome. Sleep Med 2005;6:115-21.

8. Aloia MS, Arnedt JT, Davis JD, Riggs RL, Byrd D. Neuropsycho-logical sequelae of obstructive sleep apnea-hypopnea syndrome: a critical review. J Int Neuropsychol Soc 2004;10:772-85.

9. Beebe DW, Groesz L, Wells C, Nichols A, McGee K. The neuro-psychological effects of obstructive sleep apnea: a meta-analysis of norm-referenced and case-controlled data. Sleep 2003;26:298-307.

10. Aikens JE, Caruana-Montaldo B, Vanable PA, Tadimeti L, Mendel-son WB. MMPI correlates of sleep and respiratory disturbance in obstructive sleep apnea. Sleep 1999;22:362-9.

11. Guilleminault C, Robinson A. Sleep-disordered breathing and hy-pertension: past lessons, future directions. Sleep 1997;20:806-11.

12. Sassani A, Findley LJ, Kryger M, Goldlust E, George C, David-son TM. Reducing motor-vehicle collisions, costs, and fatalities by treating obstructive sleep apnea syndrome. Sleep 2004;27:453-8.

13. Brown WD. The psychosocial aspects of obstructive sleep apnea. Semin Respir Crit Care Med 2005;26:33-43.

14. Lanfranchi P, Somers VK. Obstructive sleep apnea and vascular dis-ease. Respir Res 2001;2:315-9.

15. Beebe DW, Gozal D. Obstructive sleep apnea and the prefrontal cor-tex: towards a comprehensive model linking nocturnal upper airway obstruction to daytime cognitive and behavioral deficits. J Sleep Res 2002;11:1-16.

16. Davies CW, Crosby JH, Mullins RL, et al. Case control study of cerebrovascular damage defined by magnetic resonance imaging in patients with OSA and normal matched control subjects. Sleep 2001;24:715-20.

17. Macey PM, Henderson LA, Macey KE, et al. Brain morphology associated with obstructive sleep apnea. Am J Respir Crit Care Med 2002;166:1382-7.

18. Ashburner J, Friston KJ. Voxel-based morphometry--the methods. Neuroimage 2000;11:805-21.

19. O'Donoghue FJ, Briellmann RS, Rochford PD, et al. Cerebral struc-tural changes in severe obstructive sleep apnea. Am J Respir Crit Care Med 2005;171:1185-90.

20. Macey PM, Harper RM. OSA brain morphology differences: mag-nitude of loss approximates age-related effects. Am J Respir Crit

Care Med 2005;172:1056-7; author reply 7-8.21. Morrell MJ, McRobbie DW, Quest RA, Cummin AR, Ghiassi R,

Corfield DR. Changes in brain morphology associated with ob-structive sleep apnea. Sleep Med 2003;4:451-4.

22. Gale SD, Hopkins RO. Effects of hypoxia on the brain: neuroim-aging and neuropsychological findings following carbon monox-ide poisoning and obstructive sleep apnea. J Int Neuropsychol Soc 2004;10:60-71.

23. Blatter DD, Bigler ED, Gale SD, et al. Quantitative volumetric analysis of brain MR: normative database spanning 5 decades of life. AJNR Am J Neuroradiol 1995;16:241-51.

24. Ding J, Nieto FJ, Beauchamp NJ, Jr., et al. Sleep-disordered breath-ing and white matter disease in the brainstem in older adults. Sleep 2004;27:474-9.

25. Robbins JA, Redline S, Ervin A, Walsleben JA, Ding J, Nieto FJ. Associations of sleep-disordered breathing and cerebral changes on MRI. J Clin Sleep Med 2005;1:159-65.

26. Narayana PA. Magnetic resonance spectroscopy in the monitoring of multiple sclerosis. J Neuroimaging 2005;15:46S-57S.

27. Abbott C, Bustillo J. What have we learned from proton magnetic resonance spectroscopy about schizophrenia? A critical update. Curr Opin Psychiatry 2006;19:135-9.

28. Forton DM, Allsop JM, Cox IJ, et al. A review of cognitive impair-ment and cerebral metabolite abnormalities in patients with hepati-tis C infection. AIDS 2005;19:S53-63.

29. Larsson HB, Christiansen P, Jensen M, et al. Localized in vivo pro-ton spectroscopy in the brain of patients with multiple sclerosis. Magn Reson Med 1991;22:23-31.

30. Meyerhoff DJ, MacKay S, Constans JM, et al. Axonal injury and membrane alterations in Alzheimer’s disease suggested by in vivo proton magnetic resonance spectroscopic imaging. Ann Neurol 1994;36:40-7.

31. van der Grond J, Balm R, Kappelle LJ, Eikelboom BC, Mali WP. Cerebral metabolism of patients with stenosis or occlusion of the in-ternal carotid artery. A 1H-MR spectroscopic imaging study. Stroke 1995;26:822-8.

32. Kamba M, Suto Y, Ohta Y, Inoue Y, Matsuda E. Cerebral metabo-lism in sleep apnea. Evaluation by magnetic resonance spectros-copy. Am J Respir Crit Care Med 1997;156:296-8.

33. Kamba M, Inoue Y, Higami S, Suto Y, Ogawa T, Chen W. Cerebral metabolic impairment in patients with obstructive sleep apnoea: an independent association of obstructive sleep apnoea with white matter change. J Neurol Neurosurg Psychiatry 2001;71:334-9.

34. Bartlett DJ, Rae C, Thompson CH, et al. Hippocampal area metabo-lites relate to severity and cognitive function in obstructive sleep apnea. Sleep Med 2004;5:593-6.

35. Rae C, Digney AL, McEwan SR, Bates TC. Oral creatine monohy-drate supplementation improves brain performance: a double-blind, placebo-controlled, cross-over trial. Proc Biol Sci 2003;270:2147-50.

36. Alchanatis M, Deligiorgis N, Zias N, et al. Frontal brain lobe im-pairment in obstructive sleep apnoea: a proton MR spectroscopy study. Eur Respir J 2004;24:980-6.

37. Nofzinger EA. Neuroimaging and sleep medicine. Sleep Med Rev 2005;9(3):157-72.

38. Henderson LA, Woo MA, Macey PM, et al. Neural responses dur-ing Valsalva maneuvers in obstructive sleep apnea syndrome. J Appl Physiol 2003;94:1063-74.

39. Tiecks FP, Lam AM, Matta BF, Strebel S, Douville C, Newell DW. Effects of the valsalva maneuver on cerebral circulation in healthy adults. A transcranial Doppler Study. Stroke 1995;26:1386-92.

40. Felker GM, Cuculich PS, Gheorghiade M. The Valsalva maneuver: a bedside “biomarker” for heart failure. Am J Med 2006;119:117-22.

41. Harper RM, Macey PM, Henderson LA, et al. fMRI responses to cold pressor challenges in control and obstructive sleep apnea sub-jects. J Appl Physiol 2003;94:1583-95.



42. Macey PM, Macey KE, Henderson LA, et al. Functional magnetic resonance imaging responses to expiratory loading in obstructive sleep apnea. Respir Physiol Neurobiol 2003;138:275-90.

43. Macey KE, Macey PM, Woo MA, et al. Inspiratory loading elic-its aberrant fMRI signal changes in obstructive sleep apnea. Respir Physiol Neurobiol 2006;151:44-60.

44. Thomas RJ, Rosen BR, Stern CE, Weiss JW, Kwong KK. Func-tional imaging of working memory in obstructive sleep-disordered breathing. J Appl Physiol 2005;98:2226-34.

45. Ayalon L, Ancoli-Israel S, Klemfuss Z, Shalauta MD, Drummond SPA. Increased brain activation during verbal learning in obstruc-tive sleep apnea. Neuroimage:In Press.

46. Drummond SP, Brown GG, Gillin JC, Stricker JL, Wong EC, Bux-ton RB. Altered brain response to verbal learning following sleep deprivation. Nature 2000;403:655-7.

47. Drummond SP, Meloy MJ, Yanagi MA, Orff HJ, Brown GG. Com-pensatory recruitment after sleep deprivation and the relationship with performance. Psychiatry Res 2005;140:211-23.

48. Portas CM, Rees G, Howseman AM, Josephs O, Turner R, Frith CD. A specific role for the thalamus in mediating the interaction of attention and arousal in humans. J Neurosci 1998;18:8979-89.

49. Cabeza R, Grady CL, Nyberg L, et al. Age-related differences in neural activity during memory encoding and retrieval: a positron emission tomography study. J Neurosci 1997;17:391-400.

50. Morcom AM, Good CD, Frackowiak RS, Rugg MD. Age effects on the neural correlates of successful memory encoding. Brain 2003;126:213-29.

51. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The oc-currence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230-5.

52. Shepertycky MR, Banno K, Kryger MH. Differences between men and women in the clinical presentation of patients diagnosed with obstructive sleep apnea syndrome. Sleep 2005;28:309-14.

53. Shaw P, Greenstein D, Lerch J, et al. Intellectual ability and cortical development in children and adolescents. Nature 2006;440:676-9.

54. Alchanatis M, Zias N, Deligiorgis N, Amfilochiou A, Dionellis G, Orphanidou D. Sleep apnea-related cognitive deficits and in-telligence: an implication of cognitive reserve theory. J Sleep Res 2005;14:69-75.

55. Murray AD, Staff RT, Shenkin SD, Deary IJ, Starr JM, Whalley LJ. Brain white matter hyperintensities: relative importance of vascular risk factors in nondemented elderly people. Radiology 2005;237:251-7.

56. Viswanathan A, Chabriat H. Cerebral microhemorrhage. Stroke 2006;37:550-5.

57. Kotani K, Osaki Y, Sakane N, Adachi S, Ishimaru Y. Risk factors for silent cerebral infarction in the elderly. Arch Med Res 2004;35:522-4.

58. Kurata M, Okura T, Watanabe S, Higaki J. Association between carotid hemodynamics and asymptomatic white and gray mat-ter lesions in patients with essential hypertension. Hypertens Res 2005;28:797-803.

59. Beache GM, Herzka DA, Boxerman JL, et al. Attenuated myocar-dial vasodilator response in patients with hypertensive hypertrophy revealed by oxygenation-dependent magnetic resonance imaging. Circulation 2001;104:1214-7.

60. Wang R, Foniok T, Wamsteeker JI, et al. Transient blood pressure changes affect the functional magnetic resonance imaging detection of cerebral activation. Neuroimage 2006:Epub ahead of print.

61. Lin L, Finn L, Zhang J, Young T, Mignot E. Angiotensin-converting enzyme, sleep-disordered breathing, and hypertension. Am J Respir Crit Care Med 2004;170:1349-53.

62. Engleman HM, Martin SE, Douglas NJ. Compliance with CPAP therapy in patients with the sleep apnoea/hypopnoea syndrome. Thorax 1994;49:263-6.

63. McArdle N, Devereux G, Heidarnejad H, Engleman HM, Mackay TW, Douglas NJ. Long-term use of CPAP therapy for sleep apnea/

hypopnea syndrome. Am J Respir Crit Care Med 1999;159:1108-14.

64. Krieger J, Kurtz D, Petiau C, Sforza E, Trautmann D. Long-term compliance with CPAP therapy in obstructive sleep apnea patients and in snorers. Sleep 1996;19:S136-43.

Neuroimaging in OSA

A Review of Neuroimaging in Obstructive Sleep Apnea€¦ · Journal of Clinical Sleep Medicine, Vol. 2, No. 4, 2006 461 Obstructive sleep apnea (OSA) is a common clinical sleep dis-

Documents