Evidence for a motor and a non-motor domain in the human dentate nucleus — An fMRI study

Abstract

Dum and Strick (J. Neurophysiol. 2003; 89, 634–639) proposed a division of the cerebellar dentate nucleus into a “motor” and “non-motor” area based on anatomical data in the monkey. We asked the question whether motor and non-motor domains of the dentate can be found in humans using functional magnetic resonance imaging (fMRI). Therefore dentate activation was compared in motor and cognitive tasks. Young, healthy participants were tested in a 1.5 T MRI scanner. Data from 13 participants were included in the final analysis. A block design was used for the experimental conditions. Finger tapping of different complexities served as motor tasks, while cognitive testing included a verbal working memory and a visuospatial task. To further confirm motor-related dentate activation, a simple finger movement task was tested in a supplementary experiment using ultra-highfield (7 T) fMRI in 23 participants. For image processing, a recently developed region of interest (ROI) driven normalization method of the deep cerebellar nuclei was used.

Dorso-rostral dentate nucleus activation was associated with motor function, whereas cognitive tasks led to prominent activation of the caudal nucleus. The visuospatial task evoked activity bilaterally in the caudal dentate nucleus, whereas verbal working memory led to activation predominantly in the right caudal dentate. These findings are consistent with Dum and Strick's anatomical findings in the monkey.

Introduction

Functional brain imaging and human cerebellar lesion studies have convincingly shown that the cerebellum contributes to cognitive function, including executive, language, and visuospatial abilities (Malm et al., 1998, Silveri et al., 1998, Fink et al., 2000, Fink et al., 2001, Miall et al., 2001, Gurd et al., 2002, Schmahmann, 2004, Lee et al., 2005, Ravizza et al., 2006, Frings et al., 2006, Molinari and Leggio, 2007). There is increasing evidence of a topographic organization within the human cerebellar cortex for motor and non-motor function. Recent meta-analyses of both neuroimaging and human cerebellar lesion studies suggest that sensorimotor function relies primarily on lobules I–V of the anterior lobe, adjacent parts of lobule VI, and lobule VIII of the posterior lobe, whereas cognitive functions are supported predominantly by lobules VI and VII (Stoodley and Schmahmann, 2009, Stoodley and Schmahmann, 2010). In contrast, in the human very little is known about topographic organization within the cerebellar nuclei, the main output targets of the cerebellar cortex.

Anatomical studies in monkeys have shown a somatotopic organization within the dentate nucleus (Stanton, 1980, Thach et al., 1993). In contrast to the two body representations in the cerebellar cortex, only one body representation has been found in the dentate nucleus, raising the hypothesis that both body areas may converge on similar regions in the nucleus. These early findings have been extended by more recent work by Dum and Strick (2003). Using retrograde transneuronal transport of neurotropic viruses (for a review, see Strick et al., 2009), the authors found that areas located in the dorsal dentate project to the primary motor and premotor area (PMv), while more caudal and ventral areas are connected to prefrontal (Brodmann area (BA) 46, 9 L) and parietal areas (BA 7b) (Fig. 1A). Because the prefrontal and parietal areas are known to be involved in cognitive functions, Dum and Strick postulated that the dentate nucleus can be functionally subdivided into a more dorso-rostral motor domain and a more ventral and caudal non-motor domain with distinct output channels and differing antigenic properties (Pimenta et al., 2001). The dorso-caudal dentate, however, has also been shown to project to the anterior intraparietal area (Fig. 1B; Clower et al., 2005).

A number of functional magnetic resonance imaging (fMRI) studies in healthy human participants report activation in the dentate nucleus for both cognitive and motor tasks (Kim et al., 1994, Jueptner et al., 1997, Dimitrova et al., 2006a, Habas, 2010). Likewise, more recent resting state fMRI studies suggest that the dentate nucleus contributes to both motor and non-motor control networks (Allen et al., 2005, Habas et al., 2009). However, with one exception (Dimitrova et al., 2006a), none of these studies has made an attempt to investigate the topographic organization within the dentate nucleus.

Methodological problems are a likely reason. Blood oxygenation level dependent (BOLD) signal detection within the human cerebellar nuclei poses a difficult challenge. Due to spatial variability and the small size of the dentate, the nuclei of individual subjects reach only a maximal overlap of 70% in a standard group analysis with the widely used Montreal Neurologic Institute (MNI) 152 template (Dimitrova et al., 2006b). Thus, the chance that functionally corresponding subregions of the cerebellar nuclei superimpose across participants with common normalization methods is quite small. Worse still, there is a substantial danger that reported activations of the deep cerebellar nuclei (DCN) are due to physiological artifacts or smoothed activation from the surrounding gray matter. In the present study, we therefore employed a recently developed new normalization approach (Diedrichsen et al., 2010). We utilized the fact that the dentate nucleus, due to its high iron content, can be clearly seen on T2-weighted scans as a hypointensity. We used this information to ensure overlap of the deep cerebellar nuclei after normalization.

We asked the question whether a division into motor and non-motor areas as proposed by Dum and Strick (2003) can be found in the human dentate using 1.5 T fMRI. To address this question, we compared activation in finger tapping motor tasks of different complexities and two cognitive tasks. We expected that finger movements would activate predominantly the more rostral parts of the ipsilateral dorsal dentate, whereas the cognitive tasks would activate the caudal parts of the dentate. The cognitive conditions comprised a verbal working memory and a visuospatial task. Based on the connections of the cerebellar hemispheres with the contralateral cerebrum, the right cerebellar hemisphere is thought to support language function, and the left hemisphere visuospatial function (Fink et al., 2000, Stoodley and Schmahmann, 2009). Therefore, activity related to verbal working memory was expected primarily in the right dentate and activity related to the visuospatial task primarily in the left dentate. To further substantiate the findings of a motor area within the dentate nucleus a supplementary experiment was performed using ultrahigh-field MRI (7 T). The increased signal-to-noise ratio in the 7 T MRI field was considered helpful confirming the existence of a motor area within the dorso-rostral dentate.