Show Changes Show Changes
Edit Edit
Print Print
Recent Changes Recent Changes
Subscriptions Subscriptions
Lost and Found Lost and Found
Find References Find References
Rename Rename


9/14/2007 8:31:51 AM
12/20/2005 3:56:24 PM
List all versions List all versions
Neuroimaging in MND

Check out the formatting tips on the right for help formatting and making links.

Use the template below:

SummaryA summary of neuroimaging evidence in MND

Computerised Tomography (CT)

Computerised Tomography (CT) scans rely on a rotating X-Ray source to provide clinicians and radiologists with images up to 100 times more detailed than a standard X-Ray. In ALS they were typically used in research up until the 1990’s when Magnetic Resonance Imaging (MRI) scanners became more commonplace. In one of the first case studies of MND-Dementia in 1979, CT identified frontotemporal cerebral atrophy and enlarged ventricles (Mitsuyama and Takamiya 1979). Subsequently, an early neuropsychological investigation of ALS found that 8 of 14 patients displayed cortical atrophy on CT scan although this was neither quantified nor analysed by region (David and Gillham 1986). A later study using a larger sample size refuted this finding, however, and identified no significant differences between 28 ALS patients and 28 matched controls (Gallassi et al. 1989).

More recently, serial CT was carried out on an annual basis to investigate 22 ALS patients and again found evidence of frontotemporal cerebral atrophy (Kato et al. 1993). Of note, the involvement of the precentral gyrus became more marked in a subset of patients who required ventilation during the course of the study, with further atrophy identified in the postcentral gyrus, anterior cingulate gyrus, and corpus callosum. It is unclear whether decreased respiratory function affected the brains of these patients or whether the higher degree of atrophy reflects a more advanced stage of disease. In addition, six ALS patients developed atrophy in the medial temporal and parahippocampal gyri. The distribution of cerebral atrophy appeared similar in patients with ALS and MND-D. Furthermore, three patients with prolonged survival times (between 10 and 22 years of disease) showed little or no cerebral atrophy despite a typical pattern of ALS involvement of the upper and lower motor neurones.

Structural MRI

In neuroimaging research, CT eventually became superceded by MRI, which provides quantified, higher resolution images of deep brain structures, and does not involve exposure to a radioactive source.

Apart from its applications in research, MRI is also used during the diagnostic process in ALS to exclude the differential diagnosis of MS or spinal cord damage (Kiernan and Hudson 1994; Basak et al. 2002). As MRI scanners become more widespread they represent an attractive option; they are non-invasive, produce quantifiable data, and can be repeated safely (Leigh et al. 2002). Unfortunately, they are not suitable for anyone with metal in their body (e.g. aneurysm clips, shrapnel), or people that are claustrophobic, and they are more costly than CT. An early study using structural MRI found abnormally high T-2 weighted signal in the frontal and anterior temporal lobes (Kato et al. 1993). Additionally, three of their patients had slightly increased signal intensity in the cerebral white matter anterior to the postcentral sulcus and five had increased signal in the corticospinal tract. A later, quantitative analysis compared the cortical and white matter surface area of the precentral gyri, anterior frontal cortex, and parietal lobes in 11 ALS patients, 8 PLS patients and 49 matched controls (Kiernan and Hudson 1994). They found that whilst ALS patients showed significant (~6%) reductions in frontal lobe white matter compared to controls there was no corresponding cortical atrophy in this region.

More recently, an automated volumetric analysis was used to contrast grey and white matter densities in 16 ALS patients and 8 matched controls (Ellis et al. 2001). The authors found significantly decreased regions of grey matter in the superior frontal gyrus, medial frontal gyrus, and mid frontal gyrus amongst ALS patients, once more supporting the notion of extra-motor involvement in ALS. Furthermore, there was an increased amount of white matter found in the right inferior frontal gyrus amongst ALS patients. In analyzing the clinical features of their patients, there was a higher degree of white matter loss in the corticospinal tract amongst patients with a bulbar onset as opposed to a limb onset. Patients with a limb onset also had a lower degree of grey matter density in the right cingulate gyrus and medial frontal gyrus compared with bulbar-onset patients.

Using the same MRI technique, grey and white matter densities were examined in ALS patients with and without cognitive impairment, compared to matched controls (Abrahams et al. 2004a). Whilst there was no significant reduction in grey matter densities, white matter losses were found amongst cognitively impaired ALS patients in the motor tracts, medial temporal lobe, anterior cingulate gyrus, and medial frontal lobes.

These changes included frontotemporal association fibres adjacent to the ventricles, as well as parts of the limbic system and connections between the temporal poles. ALS patients without cognitive impairment also showed a less widespread decrease in white matter density in the corpus callosum, posterior cingulum, posterior cingulate gyrus, and the association fibres of the precuneus. The authors suggest that extra-motor degeneration in ALS patients without cognitive impairment may precede cognitive change and adds support to the hypothesis of a continuum of impairment and extra-motor change in ALS.

In a large MRI study of 72 consecutive ALS patients and 56 matched controls, Frank et al (1997) found marked neuropsychological deficits amongst the majority of the group (n=45), classified as cognitively impaired. By contrast, the unimpaired group consisted of only 15 patients. In addition to poorer performance on tasks associated with frontal lobe performance such as the D2 test, WCST, and verbal fluency, the impaired group had significantly enlarged ventricles relative to patients in the unimpaired group, an indicator of cerebral atrophy. Like most of the studies described above, no correlational analysis was undertaken between cerebral atrophy and neuropsychological test performance. As these patients represent a consecutive series it is possible that some participants from the ALS group may have had MND-D, although given the rarity of this association this is unlikely to account for the magnitude of the impairment found in the majority of their patients.

Functional Imaging

Single Photon Emission Computerised Tomography (SPECT)

Unlike traditional CT or MRI, SPECT imaging produces an image of organ function within the body through the use of radionuclide-labeled tracers injected into the body. Whilst providing a different perspective from anatomical scans, the disadvantage in this technique is a relatively poor spatial and temporal resolution, little quantifiable output, and the added complication of generating and safely using a radioactive source within the body (Leigh et al. 2002).

In ALS, SPECT has been the methodology most frequently employed in examining patients with cognitive change, such as Neary et al’s study which identified reduced tracer uptake in the frontal lobes of four patients with MND-Dementia (Neary et al. 1990). Pathological examination on two of these patients revealed a corresponding degree of cortical atrophy, spongiform change, and gliosis in the frontal lobes (see Section 2.4). SPECT analysis of a cognitively heterogeneous group of ALS patients found a similar pattern of reduced frontal lobe uptake amongst a subgroup of four patients with MND-D (Abe et al. 1993). A subgroup of two ALS patients with ‘borderline dementia’ showed a similar, if less marked pattern of reduced uptake, whilst eight ALS had reduced uptake confined to the motor area. Finally, four PMA patients (with no clinical signs of UMN involvement) showed a normal pattern of tracer uptake.

A number of subsequent studies have replicated these findings, establishing via SPECT that frontotemporal lobe involvement is widespread in ALS patients, particularly those with MND-Dementia or a predominance of UMN involvement (Abe et al. 1997; Neary et al. 1990; Portet et al. 2001; Talbot et al. 1995; Vercelletto et al. 1999; Vercelletto et al. 2003; Mantovan et al. 2003). Three similar studies have used SPECT to compare neuroimaging outcomes and neuropsychological test scores amongst subgroups of ALS patients. In one study, reduced cerebral glucose uptake was found in frontal regions amongst ALS patients with cognitive impairment but not in those without (Abe et al. 1997). Another study subdivided 23 ALS patients into cognitively impaired and unimpaired groups, and found SPECT abnormalities of a frontal nature in 3 of 11 (27%) unimpaired patients and 7 of 9 (78%) impaired patients (Portet et al. 2001). Finally, the most comprehensive investigation of cognitive function carried out SPECT imaging using 99mTc-hexamethylpropylene and a neuropsychological test battery amongst 19 ALS patients, 8 MND-D patients, 29 FTD patients, and 10 controls (Talbot et al. 1995). Neuropsychologically, ALS patients performed significantly worse than controls on tasks associated with frontal lobe functioning, although only performance on a picture sequencing task reached significance. SPECT imaging found ALS patients to have a relative reduction of rCBF in bilateral anterior and medial orbitofrontal cortex, anterior and medial frontal cortex and anterior temporal lobes bilaterally(~20-25% reduction).

A more marked and widespread pattern of impairment was found in patients with MND-D throughout the same areas as ALS patients but also the posterior frontal (~25-30% reduction), parietal and occipital lobes bilaterally (20-25% reduction). Patients with FTD had an identical pattern of reduced rCBF as MND-D with greater reduction (>30%) throughout . Unfortunately, neuropsychological testing could not be carried out in the latter two groups of patients. Whilst a useful tool in initially identifying extra-motor change in ALS, SPECT does not easily lend itself to quantifiable correlational analysis.

Positron Emission Tomography (PET)

Like SPECT, PET relies on the detection of radiation emitted by a radiolabelled tracer (such as glucose) to construct an image of tracer uptake in the brain. However, it benefits from producing quantifiable data, is more sensitive to biological changes, and provides images of a resolution superior to those provided by SPECT (Leigh et al. 2002; Turner and Leigh 2000).

The combined use of PET and MRI allows for image co-registration, so that detailed quantitative data on the function of biological systems with PET can be associated with specific regions by detailed structural images obtained from MRI (Turner and Leigh 2000). Broadly, there are two types of PET study. PET activation studies with tracers such as radiolabelled water identify changes in cerebral blood flow whilst performing a particular task. PET Ligand/Tracer studies use ligands that bind to specific receptors, (11C)-Flumazenil for GABAA receptors for example, to evaluate in detail the structural and functional integrity of neural systems.

Ligand PET

An early study used the radiolabelled glucose ligand (18F)2-Fluoro-2-deoxy-D-glucose (FDG) to assess Regional Cerebral Metabolic Rates for Glucose (rCMRGlc) in 12 ALS patients and 11 age-matched controls (Dalakas et al. 1987). They found significantly reduced global rCMRGlc of around 21% amongst ALS patients, but not in patients with PMA, who showed around a 5% reduction. The former group also had specific regions of reduced rCMRGlc in the cortex and basal ganglia. A small number of patients had follow-up scans 12 months after the initial scan.

Of these, three of four patients with UMN involvement showed further reduction of rCMRGlc associated with worsening progression of symptoms. By contrast, a PMA patient showed no significant change in rCMRGlc despite a worsening of symptoms that caused death 3 months after the second scan. In contrast to all of these findings, CT scans showed no evidence of significant atrophy. The inclusion of patients with PMA provides an interesting group for comparison (see Chapter 4), and a more detailed analysis of the differences between these patient groups was carried out by Kew et al (Kew et al. 1994b; Kew et al. 1994a)(see Section Dalakas’ study provided early in vivo evidence of extra-motor change in patients with ALS, dissociated cerebral changes in ALS and PMA, and prompted further investigations using PET.

A slightly larger study used PET with the FDG ligand amongst 18 ALS patients and 12 controls (Ludolph et al. 1992). A subgroup of 14 patients also carried out a number of neuropsychological tests. A Region-Of-Interest (ROI) analysis found significantly decreased rCMRGlc in the frontal cortex and superior occipital cortex amongst ALS patients compared to controls. In contrast to the findings of Dalakas et al, Ludolph et al did not find an association between reduced rCMRGlc and disease severity as assessed by the Norris Score.

More recently, further evidence of extra-motor involvement in ALS was provided by Lloyd et al who used the GABAA ligand (11C) Flumazenil with PET to study cerebral dysfunction (Lloyd et al. 2000). Unlike the studies of Ludolph and Dalakas, this study employed a ligand used in identifying regions of focal neuronal loss associated with, for example, epilepsy (Turner and Leigh 2000), rather than an indirect measure such as glucose uptake. Comparing 17 ALS patients with 17 age-matched controls, they found significant bilateral reductions in Flumazenil binding in the prefrontal cortex, parietal cortex, visual association cortex, and left motor/premotor cortices. Reductions were also in the left ventral and dorsolateral prefrontal cortex, Broca’s area, the right temporal cortex and right visual association cortex. In addition to adding further evidence of extra-motor involvement in ALS, the authors suggest that reduced Flumazenil binding may be associated with a loss of interneurones. Of note, they did not find reduced Flumazenil binding in the primary motor cortex, which suggests that dysfunction in this area is due to a different neuropathological process to that of extra-motor areas.

Flumazenil has also been used to show phenotypic variations in motor and extra-motor changes in different subgroups of ALS (Turner et al. 2005a)(see Chapter 6). As expected, 24 Sporadic ALS patients showed decreased flumazenil binding relative to 24 matched controls in premotor regions, motor cortex, and posterior association areas. In addition to these groups, 10 patients with the slowly progressive D90A SOD1 mutation also underwent PET scanning with the same ligand. In contrast to the predominantly motor picture of neuronal loss, decreased flumazenil binding in this subgroup was restricted to the left fronto-temporal junction and medial anterior cingulate gyrus. Furthermore, two asymptomatic participants testing positive for the D90A SOD1 gene mutation had a similar, but less marked, pattern of neuronal loss suggesting a preclinical syndrome. Flumazenil PET is discussed in more detail in Section 6.3.

The same group investigated the use of two other ligands as potential surrogate markers for ALS. The ligand (11C)-PK11195 is a proposed marker of activated microglia and therefore a correlate of inflammatory neuropathological mechanisms. Ten ALS patients were compared with 14 age-matched controls and scanned with PET (Turner et al. 2004). Significantly increased binding, representative of activated microglia, was found amongst ALS patients in the motor cortex, pons, dorsolateral prefrontal cortex, and thalamus. In addition, there was a strong correlation between (11C)-PK11195 binding in the motor cortex and clinical extent of UMN involvement, but not with disease duration or score on the ALS Functional Rating Scale Revised (ALSFRS-R) (Cedarbaum et al. 1999). This study establishes microglia as playing a role in neuronal loss both in motor and extra-motor regions in ALS, and the authors suggest that future studies may be able to identify “at risk” individuals such as asymptomatic SOD1 mutation carriers.

Finally, the ligand WAY100635, which binds to 5HT1A (‘Serotonin’) receptors was used to investigate receptor changes in 21 Sporadic ALS patients and 19 matched controls (Turner et al. 2005b)(see Section 6.4). This study found decreased uptake globally (~21%) and more specifically between 16-29% in frontotemporal regions, cingulate, and lateral precentral gyri, more so in patients positive for bulbar involvement.

In summary, ligand PET imaging studies have found extensive evidence of extra-motor cortical change in ALS, in terms of abnormal glucose utilization, changes in a proposed marker of neuronal loss, and an increased number of activated microglia in both motor and extra-motor regions. In particular, the degree of motor cortex involvement is thought to be associated with clinical signs of UMN damage such as spasticity, brisk reflexes, and increased tone. Future studies seeking to identify objective markers of UMN change such as Transcranial Magnetic Stimulation (TMS) might be useful in future applications of PET imaging.

Activation/Functional PET

A series of studies by Kew et al (Kew et al. 1993a; Kew et al. 1993b; Kew et al. 1994b; Kew et al. 1994a) used radiolabelled water to trace regional cerebral blood flow in ALS patients during the performance of a joystick movement task. In the first study (Kew et al. 1993b), 12 ALS patients and 6 matched controls underwent PET scanning both at rest and whilst performing a joystick movement (motor) task requiring them to make either ‘random’ joystick movements or predefined ‘stereotyped’ movements in one of four directions. During the stereotyped motor task, ALS patients had significantly increased rCBF in the contralateral regions of the ventral sensorimotor cortex (associated with the face and adjacent to the hand area), ventral premotor cortex, anterior insula and the ipsilateral anterior cingulate cortex. During the random motor task these areas also showed evidence of increased rCBF but to a more significant degree. The random motor task also produced decreased rCBF amongst ALS patients in the medial prefrontal cortex and left parahippocampal gyrus. A subsequent study also using radiolabelled water (Kew et al. 1993a) found that ALS patients showed significant bilateral reduction of Regional Cerebral Blood Flow (rCBF) in the sensorimotor cortex, lateral premotor cortex, the anterior cingulate cortex, superior parietal lobe, inferior parietal lobe, paracentral lobule, left anterior insula and right dorsolateral prefrontal cortex. By contrast, the occipital and temporal lobes had no areas of reduced rCBF at rest.

The authors also found that increased activation in the face area of the sensorimotor cortex was positively associated with the degree of upper limb spasticity in the patient. As no facial muscle movement was recorded during scanning, and as facial movement would normally cause bilateral activation, the authors suggest that increased rCBF in this area may reflect compensatory recruitment of adjacent cortical neurones as an adaptation to corticospinal tract degeneration. A later study (Kew et al. 1994b) reported no differences in rCBF at rest between ALS patients and PMA patients, although PMA patients did show significantly greater activation in the contralateral anterior insula. The authors suggest that this dissociation between abnormal sensorimotor cortex function in ALS patients but only isolated changes in the anterior insula amongst PMA patients may indicate that changes in the latter area are associated as an adaptation to weakness rather than corticospinal tract degeneration.

In the earliest PET study comparing neuropsychology findings with neuroimaging, Ludolph et al found that performance on a task of verbal fluency was significantly correlated with glucose metabolism across the entire cortex (Ludolph et al. 1992). As the most robust of neuropsychological findings (see Section, tests of verbal fluency have been used in subsequent scanning studies including several originating from the Institute of Psychiatry. Kew et al used radiolabelled water with PET to compare rCBF in two subgroups of ALS patients; 5 with impaired verbal fluency scores (ALSi), and 5 with unimpaired fluency scores (ALSu) during the performance of a motor task described above (Kew et al. 1993a). Across the whole cortex, ALSi patients showed significantly greater reduction in rCBF at rest than ALSu patients, particularly in the bilateral supplementary motor area, bilateral anterior cingulate cortex, and left paracentral lobule. During the activation task, ALSi patients showed significantly lower activation in the right parahippocampal gyrus, right anterior cingulate cortex, and bilateral anterior thalamic nuclear complex. In terms of specific test results, both verbal fluency and age scaled KOLT scores were found to be significantly correlated with activation in the right parahippocampal gyrus.

Another study by the same group used the same scanning technique to measure rCBF in two groups of 6 ALS patients; ALSi and ALSu (as above) as assessed by performance on the written verbal fluency task, as well as 6 age-matched controls (Abrahams et al. 1996).

For this study a verbal fluency activation paradigm was used whereby participants had to generate words beginning with a given letter. This was compared with a condition in which they repeated words. Amongst control participants, performance of the verbal fluency generation task significantly increased rCBF across the bilateral prefrontal cortex, right lateral premotor cortex, dorsolateral prefrontal cortex and anterior cingulate gyrus. Relative to these controls, ALSi patients showed significantly reduced activation across extensive cortical regions, notably the dorsolateral prefrontal cortex, premotor cortex, insular cortex, and thalamus, despite matched performance during scanning as measured by the number of words produced. Thus far, PET studies have provided researchers with the highest level of anatomical detail for correlates of cognitive change. Future studies may apply similar paradigms in the context of functional MRI imaging which is less invasive and becoming a more widespread technology than PET (Leigh et al. 2002)

Functional MRI (fMRI)

fMRI can demonstrate which brain regions are activated by cognitive or motor tasks. This relies on the fact that brain areas being used for a specific task experience a ~1-10% increase in regional blood flow approximately 6 seconds after task initiation. The iron content in deoxygenated blood affects the T2* weighted signal in a consistent manner known as the Blood Oxygen Level Dependant (BOLD) effect and can be detected by more powerful (≥ 1.5 Tesla) MRI scanners. This information is then processed by computer to produce a map of millimeter level voxels (3-D pixels) of brain activation. Using a simple motor task, one group found fMRI evidence that in ALS the loss of neurones in the motor and premotor cortices leads to the recruitment of compensatory areas to allow motor function (Konrad et al. 2002). In contrast to the typical pattern of localized contralateral activation during motor tasks, fMRI revealed ALS patients had more activation in ipsilateral motor areas, in addition to more widespread activation of the contralateral motor areas. The authors discuss these findings in terms of a compensatory cortical reorganization response to neuronal loss in the motor system. Whilst the authors do not specifically address the issue of the implications for adjacent frontal lobe regions, it would be interesting to have obtained neuropsychological data on these participants to identify any correlations between cognitive dysfunction and the degree of compensatory neurone recruitment.

A recent fMRI study of word retrieval in ALS (Abrahams et al. 2004a) compared 28 ALS patients to 18 matched controls during performance of two tasks: verbal fluency and confrontation naming. Despite similar levels of task performance from both groups, the ALS group had significantly reduced activation relative to controls in extensive regions of the prefrontal cortex (including Broca’s area), anterior cingulate, temporal and parietal lobes during the verbal fluency task. A similar, though less extensive pattern was also observed during the confrontation naming task, with reduced activation in the inferior frontal gyrus and regions of the temporal, parietal and occipital lobes. This study suggests a broad pattern of extra-motor involvement in these tasks with altered cortical function in predominantly frontal regions and establishes the viability of fMRI as a useful technique in integrating neuroimaging and cognitive testing.

How to create and edit pages

Recent Topics

Neuroimaging in MND