Functional abnormalities of the visual processing system in subjects with mild cognitive impairment: An fMRI study

Available online at www.sciencedirect.com

Psychiatry Research: Neuroimaging 163 (2008) 248 – 259 www.elsevier.com/locate/psychresns

Functional abnormalities of the visual processing system in subjects with mild cognitive impairment: An fMRI study Arun Lawrence Warren Bokdea,b,⁎, Patricia Lopez-Bayoa , Christine Bornc , Wentian Donga , Thomas Meindlc , Gerda Leinsingerc , Stefan Johannes Teipela , Frank Faltracoa , Maximilian Reiserc , Hans-Jürgen Möllera , Harald Hampela,b a

Dementia and Neuroimaging Research Section, Alzheimer Memorial Center and Geriatric Psychiatry Branch, Department of Psychiatry, Ludwig-Maximilian University, Munich, Germany b Discipline of Psychiatry, School of Medicine, Trinity College, The Adelaide and Meath Hospital incorporating The National Children's Hospital (AMiNCH), Dublin, Ireland c Institute for Clinical Radiology, Ludwig-Maximilian University, Munich, Germany Received 19 January 2007; received in revised form 11 June 2007; accepted 26 August 2007

Abstract Subjects with mild cognitive impairment (MCI) have a higher risk of developing Alzheimer's disease compared with healthy controls (HC). Sensory impairment can contribute to the severity of cognitive impairment. We measured the activation changes in the visual system between MCI and HC subjects. There were 16 MCI subjects with either amnestic MCI or multiple-domain + amnestic MCI and an HC group of 19 subjects. There were two tasks: (a) a face matching and (b) a location matching task. Brain activation was measured using functional magnetic resonance imaging. There were no differences in task performance. The HC group selectively activated the ventral and dorsal pathways during the face and location matching tasks, respectively, while the MCI group did not. The MCI group had greater activation than the HC group in the left frontal lobe during the location matching task. There were no areas of increased activation in the HC group compared with the MCI group. The MCI group, as a compensatory mechanism, activated both visual pathways and increased activation in the left frontal lobe during the location matching task compared with the healthy controls. To our knowledge, this is the first study that has examined visual processing in MCI. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Functional magnetic resonance imaging; Object matching; Location matching; Visual system; Face matching; Alzheimer's disease

1. Introduction Progression of cognitive decline to dementia is a major concern in individuals as they age. There seems to ⁎ Corresponding author. Discipline of Psychiatry, Trinity College, Trinity Centre for Health Sciences, The Adelaide and Meath Hospital, incorporating The National Children's Hospital (AMiNCH), Tallaght, Dublin 24, Ireland. Tel.: +353 1 896 4104; fax: +353 1 896 1313. E-mail address: [email protected] (A.L.W. Bokde).

be a transition period between a range of normal cognitive function and dementia, and this “transitional” period has been defined using various clinical syndromal terms such as mild cognitive impairment (MCI), incipient or preclinical dementia, and prodromal dementia. The various concepts reflect an attempt to operationalize in part the range within cognitive dysfunction that can be present in subjects that do not yet fulfill the definition of dementia. Even though amnesic MCI is more sensitive and specific for

0925-4927/$ - see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.pscychresns.2007.08.013

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

discriminating between preclinical Alzheimer's disease (AD) and healthy controls (HC) than MCI in general (Petersen et al., 2001a,b), some studies that have examined subjects suffering from multiple-domain MCI including an amnestic domain have found a higher conversion rate to AD than in groups of only amnesic MCI subjects (Bozoki et al., 2001; Palmer et al., 2003). It has been found that the sensitivity and specificity for differentiating between preclinical AD and HC were similar in measures of episodic memory (Tierney et al., 1996; Small et al., 1997; Elias et al., 2000; Backman et al., 2001), perceptual speed (Fabrigoule et al., 1998; Fox et al., 1998; Albert et al., 2001), and visuo-spatial skill (Howieson et al., 1997; Albert et al., 2001). Thus, multiple systems of the brain may be altered in the preclinical (predementia) phases of AD. As accurate visual function facilitates memory, attention and executive functions, perceptual dysfunction contributes to the severity of cognitive impairment (Cronin-Golomb et al., 1995; Rizzo et al., 2000). Some studies seem to indicate that the ventral visual pathway is more affected (Mendola et al., 1995; Rizzo et al., 2000) while others support greater dysfunction along the dorsal visual pathway (Gilmore et al., 2004). Various imaging studies have suggested that AD patients were more impaired in tasks that activated the dorsal visual pathway than in tasks that activate the ventral visual pathway (Mentis et al., 1996, 1998). Mentis et al. (1996) showed that area MT/V5 in the visual cortex was activated in HC when they attended to motion, but AD patients did not show activation in this area. Area MT/V5 receives input only from magnocellular neurons, the main neuronal population in the dorsal visual pathway. The perceptual effects were not limited to the visual system, as similar effects have been found within the auditory domain. Uhlmann et al. (1989, 1991) found that a significantly greater number of AD patients compared with HC had a hearing loss greater than 30 dB and there was a dose– response relationship between hearing loss and greater adjusted relative ratios of having dementia. Given the evidence that visual function is impaired in AD, we asked the question whether a group of MCI subjects would exhibit deficits in the visual system. To investigate the effects of visual function along both visual pathways, we developed two visual processing paradigms that preferentially activated either the ventral or dorsal pathway of the visual system. The hypothesis was that MCI subjects would exhibit greater dysfunction for tasks that recruit the dorsal pathway compared with the ventral pathway. We expected that the compensatory processes would include increased activation in the frontal lobes and in the ventral pathway. To our

249

knowledge, this is the first imaging study to investigate visual processing in both pathways in a group of MCI subjects. 2. Methods 2.1. Subjects There were 16 MCI and 19 HC subjects included in the study (demographic and neuropsychological profiles in Table 1). The MCI patients were recruited from a specialized university-based memory clinic. The clinical assessment included detailed medical history, neurological and neuropsychological examinations, and laboratory tests (routine hematology and biochemistry screen, thyroid function tests). Major systemic, psychiatric, or neurological illnesses were carefully investigated and excluded in all subjects by clinical and neurological examinations, blood testing (complete blood count, sedimentation rate, electrolytes, glucose, blood urea nitrogen, creatinine, liver-associated enzymes, cholesterol, high-density lipoprotein, triglycerides, antinuclear antibodies, rheumatoid factor, HIV, serum B12, folate, thyroid function tests, and urine analysis), and psychiatric examination. Subjects were excluded if they had cortical infarction, excessive subcortical vascular disease, space-occupying lesions, any type of dementia, depression, or other psychiatric or neurological disease. The diagnostic criteria (Petersen et al., 1999, 2001a) were (a) single memory impairment for the age and education of the subject, (b) corroboration of memory impairment by a close family member, (c) relatively preserved cognition for age, (d) no impairment in activities of daily living, and (e) no dementia. Clinical judgment was used to determine whether there was impairment in activities of daily living. The patients were systematically evaluated for the presence of affective symptoms, particularly depression; none of the MCI subjects had depression. The threshold for determining a cognitive impairment was 1.5 standard deviation (S.D.) below the age norms (Welsh et al., 1994; Berres et al., 2000) in the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) neuropsychological test battery (Morris et al., 1989). The diagnosis of the MCI subjects was established through consensus among the responsible psychiatric consultants (SJT, FF and HH). In particular, none of the MCI subjects could be classified as AD using standard clinical criteria (McKhann et al., 1984). The MCI subjects did not have the essential features of AD such as (a) memory impairment, (b) aphasia and/or apraxia, agnosia or impairment in executive function. In

250

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

Table 1 Demographic and neuropsychological characteristics of the HC and MCI groups

Number Age Education MMSE [30] Word list memory [30] Word list recall [10] Word list recognition [10] Verbal fluency Modified Boston naming test [15] Constructional praxis [11] Object matching — % correct Object matching — RT Location matching — % correct Location matching — RT

Healthy controls

MCI

8 M/11 F 66.7 (4.2) 12.8 (2.9) 29.2 (1.0) 23.9 (3.0) 8.3 (1.8) 9.9 (0.2) 24.1 (7.0) 14.6 (0.7) 10.5 (0.8) 91.7 (7.2) 1.53 (0.32) 92.9 (10.2) 1.36 (0.36)

8 M/8 F 69.9 (7.8) 13.2 (3.3) 27.2 (1.5) ⁎ 15.9 (3.5) ⁎ 3.8 (2.0) ⁎ 8.4 (1.6) ⁎⁎ 16.7 (4.3) ⁎⁎ 14.1 (1.7) 10.4 (1.8) 87.8 (11.3) 1.46 (0.32) 91.1 (8.85) 1.55 (0.40)

Values in brackets [ ] indicate maximum possible score for each specified test except verbal fluency for which a maximum score does not exist. Values are mean (S.D.) RT = response time in seconds. ⁎ Statistically significant difference at the P b 0.0001 level, df = 33. ⁎⁎ Statistically significant difference at the P b 0.001 level, df = 33.

addition, the deficits in the subject did not include significant impairment in social or occupational functioning. Thus, when the MCI subject did not have the above hallmarks of AD, dementia was excluded. The MCI subjects that were included had a single memory problem, denoted amnestic MCI (aMCI), or were subjects with a single memory problem and difficulties in verbal fluency, which are classified as multiple amnestic MCI (md + aMCI) (Petersen, 2004). The presumed neurodegenerative aetiology of these two subtypes of MCI is predominantly AD, so that MCI subjects in these two subtypes would be at high risk of being in a prodromal stage of AD (Petersen, 2004). The HC were recruited from the community, did not have an active neurological or psychiatric illness, or an illness that could affect cognitive function such as depression, and were independently functioning members of the community. The HC did not complain about cognitive problems, and there was no evidence of cognitive deficits as measured by neuropsychological testing using the CERAD (Morris et al., 1989). Subjects were excluded from the study on the basis of common MRI criteria, such as pacemaker implant, recent metallic implants or claustrophobia. All subjects had normal vision or vision corrected by use of MRcompatible eyeglasses. All subjects gave written informed consent to participate in the study after the study was explained to them. The study was performed in

accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the Faculty of Medicine of Ludwig-Maximilian University. 2.2. Stimuli and tasks The two tasks were a face and location matching task. The face matching task consisted of two faces presented simultaneously and participants were asked to decide on each trial if a pair of faces was identical or not. If they were, the subject would respond by pressing a button in the right hand. No response was required if the faces were dissimilar. The faces were grey scale stimuli where only the face was visible (Fig. 1). Each trial was 2.8 s long with an interval between pairs of faces of 0.318 s. There were eight trials per block, and there were three blocks of the task in each scan. The faces were derived from the Max Planck Institute for Biological Cybernetics database (Troje and Bulthoff, 1996). The location matching task, as shown in Fig. 1(b), consisted of two abstract images located within a smaller square. The smaller square was located within the large square. The subject had to decide if the location of the small square relative to the larger one was the same. The subject would press a button if the relative locations were identical. In the control task, the subject had to press the button every time an abstract image appeared. The images were identical to the images in the location matching task with the images always located in the center. There were

Fig. 1. Illustration of the cognitive task (a) face matching and (b) location matching.

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

four blocks of the control task, and the parameters for the presentation of the images were identical to the task of interest. The order of the face and location matching tasks was randomized across subjects in each group. At the beginning of each block there was a 7.2 s task instruction. Performance was monitored and the percentage correct and reaction times measured. Similar paradigms have been utilized previously, and have shown selective activation along the ventral and dorsal pathways for the face and location matching tasks, respectively (Haxby et al., 1994). These tasks, as well as similar tasks, have been found to reliably activate the visual pathways and are valid for assessing visual functional integrity (Corbetta et al., 1991a,b). A visual sensory paradigm assessed if there were differences in response of the blood level oxygenation dependent (BOLD) signal to a visual stimulus between the two groups. The task was a passive sensory stimulation paradigm, block design, with alternating blocks of fixation with a flashing checkerboard at 8 Hz. The blocks were 20 s long with three blocks of stimuli and four blocks of fixation. The sensory task was performed in the same session as the face and location matching task. 2.3. Scanning The imaging sequence was an interleaved T2⁎weighted echoplanar (EPI) sequence with 28 axial slices (4 mm slice thickness and slice gap = 1 mm, repetition time (TR) = 3.60 s, echo time (TE) = 60 ms, flip angle = 90°, field of view = 240 mm, and matrix = 64 × 64) and 69 volumes acquired per run (each volume was measured in 2.8 s with 0.8 s gap between volumes) on a 1.5 T Siemens Magneton Vision scanner (Erlangen, Germany). For anatomical reference in each subject, a T1-weighted sequence with 28 slices was acquired in the same orientation as the EPI sequence (TR = 620 ms, TE = 12 ms, flip angle= 90°, FOV = 240 mm, matrix = 224 × 256, rect. FOV = 7/8, effective thickness =1.25 mm), and a high resolution T1-weighted 3D Magnetization Prepared Rapid Gradient Echo (MPRAGE) structural image was acquired (TR = 11.4 ms, TE = 4.4 ms, flip angle = 8°, FOV = 270 mm, matrix = 224 × 256, rect. FOV = 7/8, effective thickness = 1.25 mm). 2.4. Data analysis The data were analyzed on an Intel Pentium III computer (San Jose, California, USA) running Linux (Red

251

Hat version 7.0, Red Hat Inc, Raleigh, North Carolina, USA) using AFNI (Cox, 1996) (http://afni.nimh.nih.gov/ afni/) and FSL (FMRIB Software Library — http://www. fmrib.ox.ac.uk/fsl). The initial step was to delete the first four volumes of each scan to remove the initial T1 magnetic transients. The remaining data were corrected for the timing differences between each slice using Fourier interpolation and then corrected for motion effects (6-parameter rigid body). Each run for each subject was analyzed using a fixed effects general linear model using FSL. Each model was composed of the regressor modeling the task of interest, the instructions, the time derivatives of the two previous regressors, and regressors for motion during the run. The task and instruction models were square wave-forms (on– off). The regressors for the task of interest and instructions were convolved with a standard double gamma hemodynamic response function. The data were smoothed (Gaussian filter at full width at half maximum = 8 × 8 × 8 mm) and high pass filtered with a cutoff at (1/100) Hz. The statistical results were normalized to the Montreal Neurological Institute/International Consortium for Brain Mapping 152 standard (MNI/ICBM). The location of the activation in the brain was done with reference to the Talairach and Tournoux template (Talaraich and Tournoux, 1988). To convert the MNI/ICBM coordinates to the Talairach and Tournoux coordinates, we utilized a non-linear transformation developed by M. Brett for transforming coordinate Table 2 Main activation peaks of object matching task compared with control task in healthy controls Region Right hemisphere Occipital lobe Fusiform gyrus Precuneus Parietal lobe Inferior parietal lobulus Frontal lobe Inferior frontal gyrus Middle frontal gyrus Anterior cingulate gyrus Left hemisphere Occipital lobe Inferior occipital gyrus Fusiform gyrus Frontal lobe Inferior frontal gyrus

Brodmann area x

y

z

Z value

19 19

36 −71 − 12 5.37 28 −60 40 5.27

40

38 −43

41 5.16

44 46 32

44 48 8

23 5.94 25 5.96 32 5.36

15 23 21

18 18 37

− 30 −92 −6 4.57 − 48 −86 − 9 4.81 − 46 −51 − 18 5.02

44

− 44

The 10 highest peaks were included in the table.

5

31 4.67

252

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

Table 3 Main activation peaks of the location matching task compared with the control task in healthy controls Region Right hemisphere Occipital lobe Middle occipital gyrus Precuneus

Brodmann area x

y

19 7 7

32 14 16

− 81 10 4.98 − 70 35 4.95 − 60 52 4.73

28

− 58 42 5.25

42

11

Parietal lobe Inferior parietal lobulus 40 Frontal lobe Inferior frontal gyrus 44 Left hemisphere Occipital lobe Inferior occipital gyrus 18 Middle occipital gyrus 19 Precuneus 7 18 Parietal lobe Inferior parietal lobulus 40

z

Z value

30 4.50

based on the locations of the peaks of activation in the HC group during the face matching task, were in the right hemisphere at (36, − 71, − 12) mm and in the left hemisphere at (− 46, − 51, − 18) mm. The changes in activation between face matching and control tasks and the changes between location matching and control tasks were obtained as a percentage change in activation. Statistical analysis was performed using SPSS (version 14, SPSS Inc, Chicago, Illinois, USA). 3. Results 3.1. Neuropsychological and behavioral performance

− 36 − 32 − 12 − 18

− 84 − 81 − 62 − 70

−9 15 47 29

4.39 5.34 4.73 5.36

− 38 − 39 31 4.47

The 10 highest peaks were included in the table.

location between both stereotactic spaces (see online at http://www.mrc-cbu.cam.ac.uk/Imaging/mnispace. html). The group statistical analyses were based on a mixed effects model with a voxelwise threshold of Z = 2.33 (P b 0.01) and was corrected for multiple comparisons at the P b 0.05 level using random field theory (Friston et al., 1994). The models for obtaining the activation maps of the functional tasks compared with the control task was the one-sample t-test, the model for the contrast between tasks (within group) was the paired t-test, and the model for the contrast between group (within task) was an unpaired t-test. The structural images were first edited to remove the non-brain tissue using BET (Smith, 2002). The EPI images were co-registered to the 28-slice T1-weighted image (7-parameter rigid body), the 28-slice T1weighted image was registered to the MPRAGE image, and the MPRAGE image was registered to the MNI/ICBM template (12 parameter). The statistical results from each subject were transformed into the MNI/ICBM space for group analysis. A region of interest (ROI) analysis was performed to examine the changes in activation in the middle of the fusiform gyrus in both tasks. The face fusiform area, an area shown to be the key for perception of face stimuli (Kanwisher et al., 1997; Haxby et al., 2001), is located in the middle fusiform gyrus. A spherical ROI with a diameter of 12 mm was located at the peak of activation within the right and left fusiform gyri of the activation maps during the face matching task. The locations,

There were statistically significant differences in the mean scores between the two groups in the MMSE, word list memory, word list recall, word list recognition, and verbal fluency subtests of the CERAD battery (see Table 1, t-test, P b 0.05, df = 33, uncorrected for multiple comparisons). In the naming and constructional praxis of the CERAD, there were no statistically significant differences. The memory impairment in the MCI group was distributed across word list memory, word list recall and word list recognition. As can be seen, the MCI group showed lower performance compared with the HC group in the verbal fluency subtest, even though the performance of the MCI group was within the normal range

Table 4 Main activation peaks of the object matching task compared with the control task in mild cognitive impaired subjects Region Right hemisphere Occipital lobe Inferior occipital lobe Fusiform gyrus Temporal lobe Inferior temporal gyrus Parietal lobe Inferior parietal lobulus Frontal lobe Inferior frontal gyrus Middle frontal gyrus Left hemisphere Occipital lobe Inferior occipital lobe Middle occipital lobe Fusiform gyrus Frontal lobe Inferior frontal gyrus

Brodmann area x

y

19 18

34 40

−92 12 4.82 −75 − 13 5.09

19

50

−74

− 1 4.54

40

40

−54

56 4.78

45 9

38 48

30 11

8 4.81 34 4.75

18 19 19

− 36 −92 − 6 4.97 − 46 −76 − 5 4.68 − 50 −70 − 12 4.86

44

− 46 15

The 10 highest peaks were included in the table.

z

Z value

25 4.88

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259 Table 5 Main activation peaks of the location matching task compared with the control task in mild cognitive impaired subjects Region Right hemisphere Occipital lobe Fusiform gyrus Middle occipital gyrus Occipital gyrus Parietal lobe Superior parietal lobulus Frontal lobe Inferior frontal gyrus Left hemisphere Occipital lobe Precuneus

x

37 19 19

54 44 34

− 59 − 14 4.21 − 85 15 4.59 − 76 30 4.36

7

32

− 54

54 4.38

Lingual gyrus

46

46

38

11 4.28

Inferior occipital gyrus Fusiform gyrus Temporal lobe Inferior temporal gyrus

Occipital gyrus 19 Temporal lobe Hippocampal gyrus 36 Parietal lobe Superior parietal 7 lobulus Frontal lobe Inferior frontal gyrus 47 Middle frontal gyrus 9 Superior frontal gyrus 6

z

Z value

Table 6 Location of statistically significant higher activation peaks in HC of (a) face matching compared with location matching and (b) location matching compared with the object matching task

Brodmann area

7

y

Region

− 18 − 77 −18 −57 − 32 − 84

48 4.41 53 4.71 28 4.35

− 18 − 22 − 21 4.85 − 32 − 59 − 50 − 54 − 32

58 4.08

27 − 1 3.49 18 − 15 3.81 7 59 3.45

The 10 highest peaks were included in the table.

(Berres et al., 2000; Satzger et al., 2001). The lower performance on the verbal fluency scale indicated that some of the MCI subjects in this group may have an additional cognitive impairment. Therefore the present group of MCI subjects all have amnesic deficits, and some of the subjects show additional cognitive impairments within the verbal domain. Accordingly, our group of MCI subjects can be classified as multiple-domain MCI with amnesia (md + aMCI). The additional cognitive impairment, however, was not severe enough to constitute the clinical diagnosis of dementia. Currently, the neuropathology in the aMCI and md + aMCI groups is hypothesized to be AD (Petersen, 2004). The correct response rate and response time in the face and location matching tasks were not statistically different between groups or between tasks (t-test at the P b 0.05 level, df = 33, see Table 1). The age and gender distributions were not statistically different. 3.2. Task activation in the healthy control group The initial analyses were to detail the location of the maximum peaks of activation of the face and location matching tasks compared with the control condition.

253

(a) Right hemisphere Occipital lobe Cuneus

Superior temporal gyrus Frontal lobe Medial frontal gyrus Anterior cingulate gyrus Inferior frontal gyrus Superior frontal gyrus Basal ganglia Thalamus Left hemisphere Occipital lobe Lingual gyrus Temporal lobe Parahippocampal gyrus Frontal lobe Inferior frontal gyrus

Middle frontal gyrus Superior frontal gyrus Anterior cingulate gyrus (b) Right hemisphere Occipital lobe Occipital gyrus Parietal lobe Precuneus

Brodmann area

x

y

z

Z value

18

20 14 16 22 36 38

− 75 − 64 − 58 − 101 − 70 − 82

8 7 −4 −5 −7 −4

2.97 3.76 3.19 4.15 4.65 4.26

37

40

− 49

− 16

5.52

20

38

30 32 31 44 22

−6 −8 −5 2 9

− 38 − 33 − 27 − 37 − 24

4.21 3.97 3.62 3.39 3.53

8 9 32

0 8 6

28 42 23

52 24 34

2.80 2.94 3.30

47

37

19

− 13

3.56

9

10 10

58 60

34 28

3.19 3.29

6

− 17

3

3.19

19

− 12

− 57

−6

4.71

28 34

− 16 − 20

− 10 1

− 13 − 10

3.65 3.50

45 46 47 9 46 10

− 53 − 44 − 55 − 40 − 55 − 48 − 22

27 30 35 29 27 48 56

2 17 9 − 10 27 18 25

3.06 3.15 2.83 3.42 3.81 2.82 3.08

32

− 14

21

30

2.67

19

28

− 76

28

3.63

7

10 14

− 47 − 68

61 49

4.34 4.62

17 31 19 18

(continued on next page)

254

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

Table 6 (continued) Region (b) Left hemisphere Parietal lobe Precuneus Superior parietal lobulus

3.5. Differences in activation between groups Brodmann area

x

7 7 7

− 10 −8 −6

y

z

− 64 − 68 − 68

Z value

47 38 55

4.38 4.62 4.50

The results of the contrast between the face matching task compared with the control condition are detailed in Table 2. The network recruited by the HC group for the location matching task compared with the control task are shown in Table 3.

Regarding this contrast, we wanted to examine whether the activation in the MCI group would be different than in the HC group. The MCI group showed higher activation in the frontal lobes compared with the HC in the location matching task, as shown in Table 7 and Fig. 2 (a contrast comparing MCI (location– control)–HC (location–control)). The increased activation within the right hemisphere was in the medial frontal gyrus while the increased activation in the left hemisphere was in the medial, middle and superior frontal gyri. There were no regions of higher activation in the HC group than in the MCI group. In the face matching task there were no areas that were statistically significantly different between groups. 3.6. Region of interest analysis

3.3. Task activation in the mild cognitive impaired group The initial analysis located the peaks of maximum activation in the face and location matching tasks, as detailed in Tables 4 and 5, respectively. 3.4. Differences in activation between tasks within groups We computed the differences in activation between the face and location matching tasks within each group. We expected to find selective activation in HC in the ventral and dorsal pathways for the face and location matching tasks, respectively. The areas of significantly greater activation in the contrast of face matching compared with location matching ((face–control)–(location–control)), as shown in Table 6a, were located in the right hemisphere in the inferior occipital gyrus, lingual gyrus, fusiform gyrus, cuneus, inferior and superior temporal gyrus, inferior, superior and medial frontal gyri, and anterior cingulate gyrus. In the left hemisphere we found greater activation in the lingual gyrus, parahippocampal gyrus, inferior, superior and middle frontal gyrus, and anterior cingulate gyrus. The activation peaks of the contrast of higher activation for location compared with face matching (that is (location–control)–(face–control)) in the HC were in the right occipital lobe, and bilateral parietal lobes (Table 6b). Within the MCI group, there were no regions that showed stronger activation in the face matching task than in the location matching task and vice versa. Thus, in the MCI group there was no selective activation.

The ROI analysis was done to further investigate the changes in activation between tasks and groups. In particular, we examined if the variance between tasks and between groups was significantly different. We examined the middle fusiform gyrus in both hemispheres. The mean level of activation in the face matching task in the right and left fusiform gyri in the HC was 0.54% (S.D. = 0.46%) and 0.52% (0.61%), respectively. In the location matching task, the

Table 7 Peaks of significantly higher activation in MCI greater than HC during location matching task Region Right hemisphere Frontal lobe Medial frontal gyrus

Left hemisphere Frontal lobe Inferior frontal gyrus Middle frontal gyrus Superior frontal gyrus

Medial frontal gyrus

Brodmann area

8 9 10

46 46/9 8 10 8 8 9 11 10 8 9 10

x

y

z

Z value

2 0 4

45 58 61

42 27 14

2.75 3.30 2.69

− 46 − 36 − 34 − 24 − 18 − 10 − 57 − 42 − 38 − 36 −8 −8 −8

46 46 28 58 45 28 23 55 64 47 31 42 55

20 27 47 25 38 52 34 14 6 3 41 33 7

3.59 4.57 3.34 3.97 3.90 3.11 4.04 3.36 3.18 2.62 2.54 2.97 2.75

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

255

Fig. 2. Greater activation for location matching in MCI than in HC subjects.

activation in the right and left fusiform gyri was 0.38% (0.35%) and 0.03% (0.81%), respectively. There were no statistically significant differences in activation in both ROIs between the face and location matching tasks. The Levene statistic was utilized to test the equality of variances in the activation in each ROI between the face and location tasks. In the right ROI, the Levene statistic was 0.365 (df1 = 1, df2 = 36, P = 0.55) and in the left ROI the Levene statistic was 0.62 (df1 = 1, df2 = 36, P = 0.81). There were no statistically significant differences in the homogeneity of variances in each ROI between tasks. In the MCI group, the level of activation in the right and left fusiform gyri during the face matching task was 0.53% (0.37%) and 0.48% (0.46%), respectively. During the location matching task, the activation in the right and left fusiform gyri was 0.28% (0.41%) and 0.34% (0.45%), respectively. There were no statistically significant differences in activation between tasks in either ROI. In the right ROI, the Levene statistic comparing the equality of variances of activation between the (face–control) tasks and (location–control) tasks was 0.03 (df = 1, df2 = 30, P = 0.87). In the left ROI, the Levene statistic was 0.03 (df1 = 1, df2 = 30, P = 0.87). There were no statistically significant differences in the homogeneity of variances in each ROI between tasks. In addition, we tested for the equality of variances for both ROIs between both groups. In contrast (HC–AD) for the face matching task, we found that there were no significant differences in the homogeneity of variances with a Levene statistic of 0.03 (df1 = 1, df2 = 33, P = 0.87) and 3.61 (df = 1, df2 = 33, P = 0.07) for the right and left fusiform gyri, respectively. Similarly in contrast (HC–AD) for the location matching task, we found no statistically significant differences between HC and MCI in the right (Levene statistic = 0.03, df1 = 1, df2 = 33, P = 0.86) and left (Levene statistic = 1.66, df1 = 1, df2 = 33, P = 0.21) ROIs.

3.7. Passive sensory task We found that both groups activated the primary visual areas of the brain during the passive stimulation paradigm. The contrast comparing differences in activation between the two groups showed no statistically significant differences (results not shown). 4. Discussion We have demonstrated that the HC group differentially activated the ventral and dorsal pathways for the face and location matching tasks, respectively. In contrast, the MCI subjects did not selectively activate the ventral and dorsal pathways. In addition, the MCI group showed higher activation in the frontal lobes compared with the HC when performing the location matching task. There were no areas of significantly greater activation in the HC group compared with the MCI group in either task. To our knowledge, this is the first time that the changes in the neural substrate underpinning visual function in a group of MCI subjects compared with HC were detailed within both visual pathways. 4.1. Selective activation of visual system in healthy controls The activation pattern for the face and location matching task differentially activated the ventral and dorsal pathways of the visual system, respectively. Comparison between the face and location matching tasks (Table 6) found regions in the visual system that modulate their activation depending on whether the task was to attend to face perception or to attend to spatial location of the objects. The results obtained with the HC were consistent with previous results (Corbetta et al., 1991a,b; Haxby et al., 1991, 1994), which showed that selective attention to objects or location increased

256

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

activation in the ventral and dorsal visual pathways, respectively. 4.2. No selective activation of the visual system in subjects with mild cognitive impairment The face and location matching tasks in the MCI subjects did not selectively activate the ventral and dorsal pathways, respectively. The lack of selective activation in the face and location matching tasks would indicate that the putative presence of AD neuropathology, even in preclinical stages, has an affect on the areas activated for performance of a visual task. In addition, to further examine the MCI data, we placed ROIs in the right and left fusiform gyri to examine if the lack of selective activation in the MCI group was due to increased variance in the activation data or reflected a lack of change in activation. In the MCI group, there was no difference in activation between the face and location matching tasks within the right and left fusiform gyri and there was equality of variance of the data between the two tasks. In addition, there was equality of variance in the HC group between the location and face matching tasks. Testing for the homogeneity of variance between the two groups revealed no statistically significant differences. Thus the lack of selective activation in the MCI group is not due to increased variance of the data compared with the HC group. 4.3. Increased activation in the location matching task in mild cognitive impaired subjects compared with healthy controls We found that there was increased activation along the dorsal pathway in the MCI group compared with the HC group in the location matching task. The increased activation was located in the frontal lobes. Thus, the MCI group in addition to non-selective activation of the ventral and dorsal visual pathways showed greater activation in the left frontal lobe during the location matching task. We hypothesized the higher activation within the left frontal lobe as a compensatory mechanism. This compensatory process could have involved various processes: (a) inefficient processing along the occipital–parietal regions, (b) the MCI subjects utilized a different strategy, or (c) the MCI subjects utilized the same strategy but recruited different brain regions, or (d) the MCI subjects were attending to information non-germane to the task. Even though there was activation along the dorsal pathway, it may have been that processing of the stimuli was inefficient compared with HC, leading to recruitment of the frontal lobe for performance of the task.

Another compensatory process that could have been active would have been utilization of a different strategy, for example, extended use in MCI subjects of the frontal lobe (higher order functions) to successfully solve the task. Another possibility is that the MCI subjects utilized the same strategy as the HC but recruited different brain regions. The lack of selective activation along the ventral and dorsal pathways in the MCI group further suggested the possibility that the MCI subjects were not selectively attending to the germane information needed for performance of the task. The compensatory mechanism may be due to the fact that the group of MCI subjects in this study was composed of individuals with md + aMCI subtype, which would have deficits in another cognitive domain in addition to an isolated, memory impairment. Thus the heterogeneity of the MCI group may play a significant role in the compensation processes that may arise. The increased activation within the frontal lobe in the location matching task compared with the HC and no between-group differences in the face matching task would support the hypothesis that the magnocellular dominated pathway has increased susceptibility to neurodegenerative damage than the ventral pathway. The work of Mentis et al. (1996, 1998) showed that there was increased vulnerability of the magnocellular cells in AD patients demonstrated by utilizing a flashing visual stimulus to activate the visual system. The present data extended these previous results to a group of MCI subjects from studies with AD patients. In addition, it extended the results by utilizing cognitive tasks that selectively activated either visual pathway in HC. The present results suggest that even when behavioral performance between groups does not differ, the neural systems that support performance may not be the same. The differences in activation pattern that we found may happen when the optimal or ideal network (as defined by the age-matched HC) is compromised by disease. In the case of the MCI subjects, the lesion or lesions that define a subject as having MCI may be “partial” in the sense that the ideal network is only partially impaired. Thus this could lead to mild or subclinical changes in performance that may go undetected unless functional brain imaging techniques are utilized to measure brain activation during performance of the task. If this is true, it would be important to find out whether the early subtle prodromal changes would be reversible or not using pharmacological or cognitive treatment strategies. Previous studies that have examined functional reorganization within the visual system have examined only the dorsal pathway (Pfefferbaum et al., 2001; Prvulovic et

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

257

al., 2002). Prvulovic et al. (2002) found in a visuo-spatial processing task that mild to moderate AD patients recruited the ventral pathway as a compensatory process for increased dysfunction within the parietal lobe. Pfefferbaum et al. (2001) in a group of alcoholics found that a visuo-spatial working memory task that activated the dorsal pathway in healthy subjects involved recruitment of areas along the ventral pathway and in the frontal lobes in the patient group. Compensatory processes were present not only in patient populations but also in groups of cognitively normal subjects that were at high risk for AD (Bookheimer et al., 2000). The functional reorganization found in these studies was heterogeneous with the functional changes dependent of the cognitive paradigm, the risk group, and the severity of the disease.

The MCI group was probably a heterogeneous group composed of: (a) those that will convert to AD in the future and (b) those that may remain classified as MCI, and (c) those subjects that may convert to other types of dementia or neurological diseases that cause cognitive dysfunction. Based on previous studies, it can be expected that from 50 to 80% will convert to AD within 5 years (Tierney et al., 1996; Bowen et al., 1997; Petersen et al., 1999; Bennett et al., 2002). Following the progression of clinical development of the subjects will allow us to better elucidate in the future the difference in activation between those MCI subjects that convert to AD and those that do not.

4.4. Heterogeneity of the MCI group

The passive stimulus task assessed if the two groups had a difference in the BOLD signal magnitude in response to a stimulus. Lack of differences between the groups would indicate that the difference in activation that we found between the MCI and HC groups is unlikely to be due to a global difference in BOLD magnitude between groups.

The MCI classification is composed of various subtypes dependent upon the composition of the main cognitive dysfunction. The most typical MCI subject is one that has memory impairment beyond what is considered to be normal within the age group and has relatively intact cognition outside of the memory domain. This type is denoted aMCI and these subjects have a high risk of converting to AD. In addition, there are other MCI subjects who have a memory impairment with an additional dysfunction beyond memory, and these subjects are denoted as multiple-domain amnestic MCI (md + aMCI). Another subgroup of MCI subjects are those that have multiple impairments outside of memory but no memory impairment — denoted as md − aMCI. A further group of subjects are those that have a single non-memory impairment in areas such as language, attention, or visuo-spatial skills. The different impairments would suggest that different neuropathological processes are present in the brain. For example, the different subtypes of MCI can be caused by different aetiologies, such as aMCI and md + aMCI have a high likelihood of converting to AD (Petersen, 2004). The hypothesized aetiologies causing the md − aMCI subtype are Lewy–Body disease and vascular dementia. The operational classification of the MCI subject into the different clinical subtypes can be an important strategy for predicting the eventual type of dementia that the MCI subject will convert to. If the MCI subject has a single non-memory domain impairment, it is presumed that the aetiology is frontal– temporal dementia or Lewy–Body dementia (Petersen, 2004). This is particularly true as long as we lack clear neurobiological measures and predictors of progression and early aetiological detection.

4.5. Hemodynamic response function

5. Conclusions The present study identified abnormalities in visual function in MCI subjects. Investigating the relationship between visual dysfunction and higher cognitive function in MCI and AD would further clarify the neurodegenerative effects on the brain. Acknowledgements This study was supported by a grant from the Volkswagen Foundation (Hannover, Germany) to A.L.W.B., S.J.T. and H.H. and by a grant from the German Competency Network on Dementias funded by the Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung), Berlin, Germany. We also thank the nursing staff at the Alzheimer Memorial Center and Geriatric Psychiatry Branch, Department of Psychiatry, LudwigMaximilian University, Munich, Germany for their assistance and support in this project. References Albert, M.S., Moss, M.B., Tanzi, R., Jones, K., 2001. Preclinical prediction of AD using neuropsychological tests. Journal of the International Neuropsychological Society 7, 631–639. Backman, L., Small, B.J., Fratiglioni, L., 2001. Stability of the preclinical episodic memory deficit in Alzheimer's disease. Brain 124, 96–102.

258

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259

Bennett, D.A., Wilson, R.S., Schneider, J.A., Evans, D.A., Beckett, L.A., Aggarwal, N.T., Barnes, L.L., Fox, J.H., Bach, J., 2002. Natural history of mild cognitive impairment in older persons. Neurology 59, 198–205. Berres, M., Monsch, A.U., Bernasconi, F., Thalmann, B., Stahelin, H.B., 2000. Normal ranges of neuropsychological tests for the diagnosis of Alzheimer's disease. Studies in Health Technology and Informatics 77, 195–199. Bookheimer, S.Y., Strojwas, M.H., Cohen, M.S., Saunders, A.M., Pericak-Vance, M.A., Mazziotta, J.C., Small, G.W., 2000. Patterns of brain activation in people at risk for Alzheimer's disease. New England Journal of Medicine 343, 450–456. Bowen, J., Teri, L., Kukull, W., McCormick, W., McCurry, S.M., Larson, E.B., 1997. Progression to dementia in patients with isolated memory loss. Lancet 349, 763–765. Bozoki, A., Giordani, B., Heidebrink, J.L., Berent, S., Foster, N.L., 2001. Mild cognitive impairments predict dementia in nondemented elderly patients with memory loss. Archives of Neurology 58, 411–416. Corbetta, M., Miezin, F.M., Dobmeyer, S., Shulman, G.L., Petersen, S.E., 1991a. Selective and divided attention during visual discriminations of shape, color, and speed: functional anatomy by positron emission tomography. Journal of Neuroscience 11, 2383–2402. Corbetta, M., Miezin, F.M., Shulman, G.L., Petersen, S.E., 1991b. Selective attention modulates extrastriate visual regions in humans during visual feature discrimination and recognition. Ciba Foundation Symposium, vol. 163, pp. 165–175. discussion 175–180. Cox, R.W., 1996. AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Computers and Biomedical Research, An International Journal 29, 162–173. Cronin-Golomb, A., Corkin, S., Growdon, J.H., 1995. Visual dysfunction predicts cognitive deficits in Alzheimer's disease. Optometry and Vision Science 72, 168–176. Elias, M.F., Beiser, A., Wolf, P.A., Au, R., White, R.F., D'Agostino, R.B., 2000. The preclinical phase of alzheimer disease: a 22-year prospective study of the Framingham Cohort. Archives of Neurology 57, 808–813. Fabrigoule, C., Rouch, I., Taberly, A., Letenneur, L., Commenges, D., Mazaux, J.M., Orgogozo, J.M., Dartigues, J.F., 1998. Cognitive process in preclinical phase of dementia. Brain 121, 135–141. Fox, N.C., Warrington, E.K., Seiffer, A.L., Agnew, S.K., Rossor, M.N., 1998. Presymptomatic cognitive deficits in individuals at risk of familial Alzheimer's disease. A longitudinal prospective study. Brain 121, 1631–1639. Friston, K.J., Worsley, K.J., Frakowiak, R.S.J., Mazziotta, J.C., Evans, A.C., 1994. Assessing the significance of focal activations using their spatial extent. Human Brain Mapping 1, 214–220. Gilmore, G.C., Morrison, S.R., Groth, K.E., 2004. Magnocellular deficit hypothesis in Alzheimer's disease. In: Cronin-Colomb, A., Hof, P.R. (Eds.), Vision in Alzheimer's disease, vol. 34. Karger, Basel, pp. 173–198. Haxby, J.V., Grady, C.L., Horwitz, B., Ungerleider, L.G., Mishkin, M., Carson, R.E., Herscovitch, P., Schapiro, M.B., Rapoport, S.I., 1991. Dissociation of object and spatial visual processing pathways in human extrastriate cortex. Proceedings of the National Academy of Sciences of the United States of America 88, 1621–1625. Haxby, J.V., Horwitz, B., Ungerleider, L.G., Maisog, J.M., Pietrini, P., Grady, C.L., 1994. The functional organization of human extrastriate cortex: a PET-rCBF study of selective attention to faces and locations. Journal of Neuroscience 14, 6336–6353. Haxby, J.V., Gobbini, M.I., Furey, M.L., Ishai, A., Schouten, J.L., Pietrini, P., 2001. Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science 293, 2425–2430.

Howieson, D.B., Dame, A., Camicioli, R., Sexton, G., Payami, H., Kaye, J.A., 1997. Cognitive markers preceding Alzheimer's dementia in the healthy oldest old. Journal of the American Geriatrics Society 45, 584–589. Kanwisher, N., McDermott, J., Chun, M.M., 1997. The Fusiform Face Area: a module in human extrastriate cortex specialized for face perception. Journal of Neuroscience 17, 4302–4311. McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., Stadlan, E.M., 1984. Clinical diagnosis of Alzheimer's disease: report of the NINCDS–ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 34, 939–944. Mendola, J.D., Cronin-Golomb, A., Corkin, S., Growdon, J.H., 1995. Prevalence of visual deficits in Alzheimer's disease. Optometry and Vision Science 72, 155–167. Mentis, M.J., Horwitz, B., Grady, C.L., Alexander, G.E., VanMeter, J.W., Maisog, J.M., Pietrini, P., Schapiro, M.B., Rapoport, S.I., 1996. Visual cortical dysfunction in Alzheimer's disease evaluated with a temporally graded “stress test” during PET. American Journal of Psychiatry 153, 32–40. Mentis, M.J., Alexander, G.E., Krasuski, J., Pietrini, P., Furey, M.L., Schapiro, M.B., Rapoport, S.I., 1998. Increasing required neural response to expose abnormal brain function in mild versus moderate or severe Alzheimer's disease: PET study using parametric visual stimulation. American Journal of Psychiatry 155, 785–794. Morris, J.C., Heyman, A., Mohs, R.C., Hughes, J.P., van Belle, G., Fillenbaum, G., Mellits, E.D., Clark, C., 1989. The Consortium to Establish a Registry for Alzheimer's Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer's disease. Neurology 39, 1159–1165. Palmer, K., Backman, L., Winblad, B., Fratiglioni, L., 2003. Detection of Alzheimer's disease and dementia in the preclinical phase: population based cohort study. British Medical Journal 326, 245. Petersen, R.C., 2004. Mild cognitive impairment as a diagnostic utility. Journal of Internal Medicine 256, 183–194. Petersen, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tangalos, E.G., Kokmen, E., 1999. Mild cognitive impairment: clinical characterization and outcome. Archives of Neurology 56, 303–308. Petersen, R.C., Doody, R., Kurz, A., Mohs, R.C., Morris, J.C., Rabins, P.V., Ritchie, K., Rossor, M., Thal, L., Winblad, B., 2001a. Current concepts in mild cognitive impairment. Archives of Neurology 58, 1985–1992. Petersen, R.C., Stevens, J.C., Ganguli, M., Tangalos, E.G., Cummings, J.L., DeKosky, S.T., 2001b. Practice parameter: early detection of dementia: mild cognitive impairment (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 56, 1133–1142. Pfefferbaum, A., Desmond, J.E., Galloway, C., Menon, V., Glover, G.H., Sullivan, E.V., 2001. Reorganization of frontal systems used by alcoholics for spatial working memory: an fMRI study. Neuroimage 14, 7–20. Prvulovic, D., Hubl, D., Sack, A.T., Melillo, L., Maurer, K., Frolich, L., Lanfermann, H., Zanella, F.E., Goebel, R., Linden, D.E., Dierks, T., 2002. Functional imaging of visuospatial processing in Alzheimer's disease. Neuroimage 17, 1403–1414. Rizzo, M., Anderson, S.W., Dawson, J., Nawrot, M., 2000. Vision and cognition in Alzheimer's disease. Neuropsychologia 38, 1157–1169. Satzger, W., Hampel, H., Padberg, F., Bürger, K., Nolde, T., Ingrassia, G., Engel, R.R., 2001. Zur praktischen Anwendung der CERADtestbatterie als neuropsychologisches Demenzscreening. Nervenartz 71, 196–203.

A.L.W. Bokde et al. / Psychiatry Research: Neuroimaging 163 (2008) 248–259 Small, B.J., Herlitz, A., Fratiglioni, L., Almkvist, O., Backman, L., 1997. Cognitive predictors of incident Alzheimer's disease: a prospective longitudinal study. Neuropsychology 11, 413–420. Smith, S.M., 2002. Fast robust automated brain extraction. Human Brain Mapping 17, 143–155. Talaraich, J., Tournoux, P., 1988. Co-Planar Stereotaxic Atlas of the Human Brain. Thieme Medical, New York. Tierney, M.C., Szalai, J.P., Snow, W.G., Fisher, R.H., Nores, A., Nadon, G., Dunn, E., St George-Hyslop, P.H., 1996. Prediction of probable Alzheimer's disease in memory-impaired patients: a prospective longitudinal study. Neurology 46, 661–665. Troje, N.F., Bulthoff, H.H., 1996. Face recognition under varying poses: the role of texture and shape. Vision Research 36, 1761–1771.

259

Uhlmann, R.F., Larson, E.B., Rees, T.S., Koepsell, T.D., Duckert, L.G., 1989. Relationship of hearing impairment to dementia and cognitive dysfunction in older adults. Journal of the American Medical Association 261, 1916–1919. Uhlmann, R.F., Larson, E.B., Koepsell, T.D., Rees, T.S., Duckert, L.G., 1991. Visual impairment and cognitive dysfunction in Alzheimer's disease. Journal of General Internal Medicine 6, 126–132. Welsh, K.A., Butters, N., Mohs, R.C., Beekly, D., Edland, S., Fillenbaum, G., Heyman, A., 1994. The Consortium to Establish a Registry for Alzheimer's Disease (CERAD). Part V. A normative study of the neuropsychological battery. Neurology 44, 609–614.