Sleep Medicine 6 (2005) 531–536 www.elsevier.com/locate/sleep
Original article
Voxel-based morphometry in narcolepsy Christian Brenneisa,*, Elisabeth Brandauera, Birgit Frauschera, Michael Schockeb, Thomas Triebb, Werner Poewea, Birgit Ho¨gla a
Department of Neurology, Medical University Innsbruck, Anichstr. 35, 6020 Innsbruck, Austria Department of Radiology, Medical University Innsbruck, Anichstr. 35, 6020 Innsbruck, Austria
b
Received 11 November 2004; received in revised form 15 February 2005; accepted 4 March 2005 Available online 1 July 2005
Abstract Background and purpose: Voxel-based morphometry (VBM) is a magnetic resonance imaging (MRI)-based indirect volumetry, which allows the investigation of the entire brain without restriction of predefined regions-of-interest. Recent studies using this technique reported controversial results in patients with narcolepsy. Patients and methods: In this study, 12 patients with narcolepsy according to the criteria of the International Classification of Sleep Disorders were compared to 12 age-matched controls with normal MR images using VBM. Sex and global differences in voxel intensities were used as confounding covariates. Results: Significant gray matter loss was found in the right prefrontal and frontomesial cortex of patients with narcolepsy. White matter comparison revealed no significant changes in patients or controls. The comparison of cerebrospinal fluid partition detected an enlargement of subarachnoidal space of controls close to the prefrontal cortex. Conclusions: The volume reduction of gray matter in narcoleptic patients could indicate a disease-related atrophy pattern. However, the results of VBM studies in narcolepsy are contradictory. A possible systematic bias due to inhomogeneous patient groups, stimulant medication history or pre-statistical image processing must be considered. We suggest that studies with drug-naive patients and/or region-ofinterest-based volumetric studies should be performed in areas defined by VBM. q 2005 Elsevier B.V. All rights reserved. Keywords: Narcolepsy; Voxel-based morphometry; Prefrontal; Frontomesial; MSLT
1. Introduction Narcolepsy is a life-long disabling disorder characterized by excessive daytime sleepiness and additional symptoms such as cataplexy, hypnagogic hallucinations, and sleep paralysis according to the International Classification of Sleep Disorders (ICSD) [1]. The reduction or absence of hypocretin/orexin neurons in the lateral hypothalamic region seems to play an important role in the pathophysiology of narcolepsy [2–4]. Hypocretin/orexin containing neurons have widespread projections to brainstem nuclei, thalamus and additional targets in the cerebral cortex [5–7]. A high association of * Corresponding author. Tel.: C43 512 504 81810; fax: C43 512 504 24288. E-mail address:
[email protected] (C. Brenneis).
1389-9457/$ - see front matter q 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.sleep.2005.03.015
narcolepsy with HLA DRB 1*1501 and DQB 1*0602 genotypes was shown [8]. HLA-positivity, however, is not necessary for diagnosis [9]. Voxel-based morphometry is an indirect volumetric method based on MR images [10] which revealed cortical atrophy in several neurodegenerative diseases [11–13]. Controversial findings have been reported in narcolepsy using this technique. Hypothalamic gray matter changes [14] as well as inferior temporal and frontal atrophy [15] were obtained by two studies whereas Overeem et al. [16] recently found no significant atrophy in the gray or white matter of patients with narcolepsy. Our study was designed to determine morphometric changes in brains of patients with narcolepsy. A preliminary analysis of these data with a standard VBM protocol revealed atrophy in frontomesial areas and an additional small area in the parietal cortex (published as abstract [17]). For this publication we used an optimized VBM protocol for pre-statistical image processing [18].
532
C. Brenneis et al. / Sleep Medicine 6 (2005) 531–536
2. Subjects Twelve patients (eight males, four females) with narcolepsy and 12 healthy age-matched controls (10 males, two females) with normal MR images were included. Mean age was 35.0G8.4 years in controls and 35.8G13.2 years in patients. Narcolepsy was diagnosed according to ICSD criteria [1]. Two patients refused polysomnography, and one had only one sleep onset REM period in five nap opportunities of the multiple sleep latency test (MSLT) while on treatment with low-dose clomipramine. In these patients, the diagnosis was made on the basis of daytime sleepiness plus cataplexy. One patient had no cataplexy, but MSLT findings were unambiguous. Clinical characteristics, HLA type, MSLT findings and current medication of patients are shown in Table 1. All subjects gave written informed consent before MR investigation. 2.1. Magnetic resonance protocol Images of all subjects were acquired on a 1.5 T MR Scanner (Magnetom Vision, Siemens). A sagittal T1weighted FLASH 3D sequence was performed with a repetition time (TR) of 9.7 ms, an echotime (TE) of 4 ms, a slice thickness of 1.5 mm, a matrix of 256!256, and a field of view of 230 mm. 2.2. Preprocessing and statistics Statistical parametric mapping (SPM 99) (The Wellcome Department of Cognitive Neurology, Institute of Neurology, Queen Square, London, UK) implemented in Matlab 5.3 (Mathworks, Inc., Sherborn, MA, USA) was used for prestatistical image processing and statistical analysis. To avoid potential bias from the scanner and normalization process, a customized template was created
including all T1-weighted images of the participating subjects. Each image was first spatially normalized into standardized MNI (Montreal Neurological Institute) space using a 12-parameter affine transformation and a nonlinear normalization by 7!8!7 basis functions. Following normalization, a mean image was created which was smoothed with an 8 mm full width at half maximum (FWHM) isotropic Gaussian kernel. The images of the subjects were warped to match this customized template applying a 12-parameter affine transformation and a nonlinear spatial normalization using discrete cosine (7!8!7) basis functions. Following reslicing onto a voxel size of 1!1!1 mm to minimize partial volume effects, images were segmented into gray and white matter as well as cerebrospinal fluid compartments. A correction of intensity nonuniformity as implemented in SPM99 was added to compensate for variations in tissue density values caused by the head position relative to the coil in the scanner. Several missegmented areas (i.e. dural venous sinus) were removed by multiplying the gray matter partition with a binary mask, which only contained the brain volume (brain extraction—a function of SPM99). A modulation of the segmented partitions was performed to compensate for volume changes in nonlinear spatial normalization by multiplying the voxel densities with the Jacobian determinants [18]. Finally, the modulated gray matter, white matter and cerebrospinal fluid partitions were convolved with a Gaussian kernel filter of 10!10!10 mm FWHM in order to render the data more normally distributed and to compensate for inexact spatial normalization [10]. The normalized, segmented, modulated and smoothed data were statistically tested using the general linear model based on the Gaussian field theory. Sex and global differences in voxel intensities were used as confounding covariates. Comparisons at group level were performed
Table 1 Clinical characteristics of patients with narcolepsy N
1 2 3 4 5 6 7 8 9 10 11 12
Age (years)
DD (years)
Cataplexy
HH
43 34 33 31 44 22 72 22 34 33 32 29
14 16 12 15 14 6 55 9 11 3 20 5
C C C K C C C C C C C C
C C K K K K C C C C C C
SP
C C K K K C K K C C C K
HLA type
MSLT
Medication
DRB 1*15
DQB 1*06
Sleep latency min
Number of SOREM
1*1501 n.a. 1*15 1*1501 1*1501 1*15 1*1501 1*15 1*15 1*15 1*15 n.a.
1*0602 n.a. 1*06 1*0602 1*0602 1*06 1*0602 1*06 1*06 1*06 1*06 n.a.
n.a. 3.5 1.6 5.2 n.a. 3 4.3 3.4 8.5 6 1.4 6.4
n.a. 4/5 4/5 2/5 n.a. 3/4 2/5 5/5 3/5 1/5 5/5 5/5
Modafinil 200 mg, protriptyline 20 mg Modafinil 100 mg Modafinil 100 mg, climiptramine 10 mg Methylphenidat 20 mg Modafinil 100 mg, climiptramine 50 mg none Methylphenidate 10 mg Modafinil 100 mg, climiptramine 25 mg Modafinil 200 mg, climiptramine 20 mg Climiptramine 50 mg Modafinil 100 mg none
DD, disease duration; HH, hypnagogic hallucination; SP, sleep paralysis; MSLT, multiple sleep latency test; SOREM, sleep onset REM; n.a., not applicable.
C. Brenneis et al. / Sleep Medicine 6 (2005) 531–536
between the tissue partitions of patients and controls as well as vice versa. The significance level was set at P!0.05 corrected for multiple comparisons across the entire brain volume.
3. Results Compared to controls, patients showed a cluster of gray matter loss in the middle frontal gyrus of the right prefrontal cortex corresponding to Brodmann area (BA) 9 (Talairach coordinates xZ46, yZ2, zZ38). An additional cluster was found in the frontomesial region corresponding to BA 10/32 (Talairach coordinates xZK3, yZ61, zZ0). Fig. 1 illustrates the gray matter loss in patients. No significant clusters were observed in the temporal or hypothalamic region. Comparing the white matter partitions of patients to controls and vice versa, no significant morphometric changes were found. VBM of subarachnoidal cavities revealed a significant cluster of enlargement in the right prefrontal region (Talairach coordinates xZ57, yZ1, zZ 52) of controls when compared to patients and no significant enlargement in the patient group.
4. Discussion In the present VBM study we found gray matter loss in the right dorsolateral prefrontal cortex (DLPFC) and the frontomesial region of patients with narcolepsy. These regions are thought to participate in many higher cognitive
533
functions such as working memory and executive processes [19,20]. Several authors have reported attention-related deficits in narcolepsy, and cognitive deficits independent of excessive daytime sleepiness have been discussed [21–26]. If we interpret our observation as a disease-specific atrophy pattern it could be related with attentional deficits and/or cognitive impairments in patients with narcolepsy. Three previous VBM studies in patients with narcolepsy revealed controversial results, ranging from gray matter loss in multiple cortical (predominantly inferior temporal and inferior frontal) areas [15] and specific subcortical areas (hypothalamus and accumbens nucleus) [14], to no morphometric changes at all [16]. Clinical characteristics of the patient collectives and results of our and the previous VBM studies are summarized in Table 2. These discrepancies need to be discussed in the light of methodological and clinical aspects, especially different data processing, inhomogeneous patient groups and pretreatment. In a preliminary analysis of our data with the standard VBM protocol we found cortical atrophy in the frontomesial cortex similar to the optimized protocol but additional areas in the parietal cortex [17]. This indicates that pre-statistical processing influences the results. One critical step in VBM is tissue segmentation, specifically in the brainstem where small gray matter areas are embedded in white matter tissue. To check for systematic bias, we additionally performed a statistical cross-check between controls and patients. If atrophy of gray matter is confounded by segmentation bias, one would expect atrophy of white matter in the corresponding region of controls. The high-density contrast between gray matter and
Fig. 1. Clusters of matter loss in the right prefrontal (sagittal, coronal and axial view) and fronto-mesial (axial view) cortex of patients with narcolepsy. x,y,z refer to Talairach coordinates.
534
Table 2 Clinical characteristics of patients with narcolepsy and VBM findings in the different publications Age
Cataplexy
MSLT sleep latency (S1) min
MSLT number of SOREM
HLA
Disease duration
Hypocretin-1 measurement
Treatment
Stimulants
Anticataplectic medication
GHB
VBM findings
Brenneis et al.
12 (8m, 4f)
35.8G13.2 (22–72)
11
4.3G2.2
3.4G1.4
DRB1*15 and DQB1*06: 11C, 1 not availablea
15.0G13.5 (3–55)
not done
10/12
2 methylphenidate 7 modafinil
6 tricyclics
0
Overeem et al. (2003)
15 (7m, 8f)
44.7G14.3 (21–70)
15
2.9 G 2.2b
2.9 G 1.2b
DQB1*0602: 15C
19.2 G 15. 0b
done
13/15
8
Kaufmann et al. (2002)
12 (6m, 6f)
36.9G15.8 (22–65)
12
n.i.
O1
DR2:11C, 1K
12b
n.i.
6/12
1 methyl5 triphenidate 6 cyclics modafinil 3 1 ssri mazindol 6 psychostimulants and tricyclic antidepressants not further specified
Draganski et al. (2002)
29 (12m,17f)
39.7G11.3 (n.i.)
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
Regional atrophy of right dorsolateral prefrontal cortex and frontomesial region of patients. Significant enlargement of subarachnoidal cavities in the right prefrontal region of controls. No significant difference between global or regional gray or white matter volume. 4% global gray matter volume loss. 19 affected areas, mainly inferior temporal and inferior frontal areas Bilat. decreases in hypothalamic gray matter volume, right accumbens nucleus, cerebellar vermis, superior temporal gyrus
n.i.
MSLT, multiple sleep latency test; SOREM, sleep onset REM; GHB, gammahydroxybutyrate; VBM, voxel-based morphometry; n.i., not indicated. a In most patients high resolution HLA typing was performed (HLA DRB1*1501 and HLA DQB1*0602), but low resolution HLA typing is reported uniformly. b Recalculated from data given in the paper. n.i., not indicated.
0
C. Brenneis et al. / Sleep Medicine 6 (2005) 531–536
n (sex)
C. Brenneis et al. / Sleep Medicine 6 (2005) 531–536
cerebrospinal fluid usually allows an appropriate classification. In our study such cross-check revealed no significant white matter changes in controls but an enlargement of subarachnoidal space close to the right prefrontal cortex. Based on the high contrast, a segmentation bias is unlikely, but cannot definitely be excluded. None of the previous studies, which found significant changes in narcolepsy, reported whether or not a vice versa comparison was performed. Inhomogeneous patient groups and small sample size could also have contributed to conflicting results. Narcolepsy seems to be a heterogeneous disease. It is well conceivable that different patient characteristics, such as cataplexy, hypocretin levels or HLA types, may affect morphometry determined by VBM [8,27]. Clinically, narcolepsy with typical cataplexy is distinguished from narcolepsy with atypical or without cataplexy [2]. All but one of our patients had typical cataplexy, but the disease duration varied from 3 to 55 years and hypocretin-1 levels were not determined. The most homogeneous patient group, characterized by undetectable hypocretin-1 levels in the cerebrospinal fluid, was studied by Overeem et al. [16]. In the largest study, details about clinical characteristics and treatment of the patients are not reported [14]. Medication, specifically long-term stimulant treatment, influences the brain metabolism and could also affect brain morphology. PET studies have shown that amphetamine decreased frontal glucose metabolism in healthy subjects [28], and a metamphetamine challenge induces prefrontal dopamine release in healthy subjects [29]. In a functional MR study [30] amphetamine affected blood flow in the right prefrontal cortex, and subjects with high working memory capacity at baseline deteriorated in cognitive tasks. Furthermore, young amphetamine-dependent subjects had significant reduction of temporal lobe volume [31]. However, in none of the previous VBM studies, have duration of psychostimulant treatment and details of substance and dose history been reported. Further volumetric studies should be performed in drug-naive narcoleptic patients or be tightly controlled for medication history. To summarize, VBM findings are not conclusive in patients with narcolepsy to date. The right prefrontal and frontomesial cortex gray matter loss found in our study might indicate a disease-related primary atrophy and/or degeneration of hypothalamo-cortical projections, but the results should be considered preliminary. To clarify the controversial results of the existing VMB studies in narcolepsy, we suggest (1) that conventional region-ofinterest-based studies based on VBM determined regions be assessed, (2) that a statistical cross-check of white and gray matter be performed (3) that highly homogenous patient samples (regarding clinical symptoms, HLA, and hypocretin/orexin from cerebrospinal fluid) be examined, and (4) that medication history and dose be paid specific attention.
535
References [1] Diagnostic Classification Steering Committee. International classification of sleep disorders: revised diagnostic and coding manual. Rochester, MN: American Sleep Disorders Association; 1997. [2] Mignot E, Lammers GJ, Ripley B, et al. The role of cerebrospinal fluid hypocretin measurement in the diagnosis of narcolepsy and other hypersomnias. Arch Neurol 2002;59:1553–62. [3] Nishino S, Ripley B, Overeem S, Lammers GJ, Mignot E. Hypocretin (orexin) deficiency in human narcolepsy. Lancet 2000; 355:39–40. [4] Thannickal TC, Moore RY, Nienhuis R, et al. Reduced number of hypocretin neurons in human narcolepsy. Neuron 2000;27:469–74. [5] Peyron C, Tighe DK, van den Pol AN, et al. Neurons containing hypocretin (orexin) project to multiple neuronal systems. J Neurosci 1998;18:9996–10015. [6] Siegel JM, Nienhuis R, Gulyani S, et al. Neuronal degeneration in canine narcolepsy. J Neurosci 1999;19:248–57. [7] van den Pol AN. Narcolepsy: a neurodegenerative disease of the hypocretin system? Neuron 2000;27:415–8. [8] Mignot E, Lin L, Rogers W, et al. Complex HLA-DR and -DQ interactions confer risk of narcolepsy–cataplexy in three ethnic groups. Am J Hum Genet 2001;68:686–99. [9] Robinson A, Guilleminault C. Narcolepsy. In: Chokroverty S, editor. Sleep disorders medicine. Basic science, technical considerations, and clinical aspects. Boston: Butterworth Heinemann; 1999. p. 427–40. [10] Ashburner J, Friston KJ. Voxel-based morphometry—the methods. Neuroimage 2000;11:805–21. [11] Brenneis C, Bosch SM, Schocke M, Wenning GK, Poewe W. Atrophy pattern in SCA2 determined by voxel-based morphometry. Neuroreport 2003;14:1799–802. [12] Brenneis C, Seppi K, Schocke MF, et al. Voxel-based morphometry detects cortical atrophy in the Parkinson variant of multiple system atrophy. Mov Disord 2003;18:1132–8. [13] Brenneis C, Seppi K, Schocke M, et al. Voxel based morphometry reveals a distinct pattern of frontal atrophy in progressive supranuclear palsy. J Neurol Neurosurg Psychiatry 2004;75:246–9. [14] Draganski B, Geisler P, Hajak G, et al. Hypothalamic gray matter changes in narcoleptic patients. Nat Med 2002;8:1186–8. [15] Kaufmann C, Schuld A, Pollmacher T, Auer DP. Reduced cortical gray matter in narcolepsy: preliminary findings with voxel-based morphometry. Neurology 2002;58:1852–5. [16] Overeem S, Steens SC, Good CD, et al. Voxel-based morphometry in hypocretin-deficient narcolepsy. Sleep 2003;26:44–6. [17] Brandauer E, Brenneis C, Frauscher B, et al. Voxel-based morphometryin Narcolepsy—a controversial issue. Sleep 2003;26: A286. [18] Good CD, Johnsrude IS, Ashburner J, et al. A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 2001;14:21–36. [19] Curtis CE, D’Esposito M. Persistent activity in the prefrontal cortex during working memory. Trends Cogn Sci 2003;7:415–23. [20] Smith EE, Jonides J. Storage and executive processes in the frontal lobes. Science 1999;283:1657–61. [21] Aguirre M, Broughton R, Stuss D. Does memory impairment exist in narcolepsy–cataplexy? J Clin Exp Neuropsychol 1985;7:14–24. [22] Fulda S, Schulz H. Cognitive dysfunction in sleep disorders. Sleep Med Rev 2001;5:423–45. [23] Henry GK, Satz P, Heilbronner RL. Evidence of a perceptualencoding deficit in narcolepsy? Sleep 1993;16:123–7. [24] Hood B, Bruck D. Sleepiness and performance in narcolepsy. J Sleep Res 1996;5:128–34. [25] Rieger M, Mayer G, Gauggel S. Attention deficits in patients with narcolepsy. Sleep 2003;26:36–43.
536
C. Brenneis et al. / Sleep Medicine 6 (2005) 531–536
[26] Rogers AE, Rosenberg RS. Tests of memory in narcoleptics. Sleep 1990;13:42–52. [27] Dauvilliers Y, Bazin M, Ondze B, et al. Severity of narcolepsy among French of different ethnic origins (south of France and Martinique). Sleep 2002;25:50–5. [28] Wolkin A, Angrist B, Wolf A, et al. Effects of amphetamine on local cerebral metabolism in normal and schizophrenic subjects as determined by positron emission tomography. Psychopharmacology 1987;92:241–6.
[29] Piccini P, Pavese N, Brooks DJ. Endogenous dopamine release after pharmacological challenges in Parkinson’s disease. Ann Neurol 2003; 53:647–53. [30] Mattay VS, Callicott JH, Bertolino A, et al. Effects of dextroamphetamine on cognitive performance and cortical activation. Neuroimage 2000;12:268–75. [31] Bartzokis G, Beckson M, Lu PH, et al. Age-related brain volume reductions in amphetamine and cocaine addicts and normal controls: implications for addiction research. Psychiatry Res 2000;98:93–102.