Segregated fronto-cerebellar circuits revealed by intrinsic functional connectivity

Abstract

Multiple, segregated fronto-cerebellar circuits have been characterized in nonhuman primates using transneuronal tracing techniques including those that target prefrontal areas. Here, we used functional connectivity MRI (fcMRI) in humans (n = 40) to identify 4 topographically distinct fronto-cerebellar circuits that target 1) motor cortex, 2) dorsolateral prefrontal cortex, 3) medial prefrontal cortex, and 4) anterior prefrontal cortex. All 4 circuits were replicated and dissociated in an independent data set (n = 40). Direct comparison of right- and left-seeded frontal regions revealed contralateral lateralization in the cerebellum for each of the segregated circuits. The presence of circuits that involve prefrontal regions confirms that the cerebellum participates in networks important to cognition including a specific fronto-cerebellar circuit that interacts with the default network. Overall, the extent of the cerebellum associated with prefrontal cortex included a large portion of the posterior hemispheres consistent with a prominent role of the cerebellum in nonmotor functions. We conclude by providing a provisional map of the topography of the cerebellum based on functional correlations with the frontal cortex.

Figure 1.

Motor and prefrontal cortex project to distinct, preferentially contralateral regions of the cerebellum. Correlation maps for motor and prefrontal seed regions are displayed overlaid on the participants’ averaged T1 structural scan. (A) Bilateral spherical seed regions in MOT (MOT coordinates: ±42, −24, 60) correlate with regions in lobules IV–VI in the anterior cerebellum and with VIIIB in ventral aspects. (B) Bilateral seed regions in DLPFC (DLPFC coordinates: ±42, 16, 36) correlate with distinct regions in Crus I and Crus II in the posterior cerebellum. In each map, red corresponds to preferentially greater correlations with seed regions in the left hemisphere and blue corresponds to preferentially greater correlations with seed regions in the right hemisphere. Maps are at a threshold of z(r) > 0.1. All image sections and atlas coordinates are referenced to the MNI coordinate system (Evans et al. 1993). Left is displayed on the left.

Figure 2.

Projections from the cerebellum form closed-loop circuits. Regions in the anterior and posterior cerebellar hemispheres correlate with distinct, nonoverlapping cerebral networks. (A) Regions correlated with CBM_MOT (lobule V) are restricted to the MOT in the frontal lobe, whereas regions correlated with CBM_DLPFC (Crus I) include lateral dorsal, ventral as well as medial PFC. Note that the CBM_MOT-correlated region at the base of the temporal lobe on the medial view is most likely actually correlated activation in the cerebellum that has “spilled over” into the cerebral cortex because of the cortical inflation and does not actually reflect correlations in the temporal lobe. Maps are at a threshold of z(r) > 0.1. The volumes are projected onto the left hemisphere cortical surface of the PALS atlas (Van Essen 2005). The right hemisphere produces indistinguishable results. Borders reflect approximate borders of relevant Brodmann areas encompassing the prefrontal cortex and MOT (see Fig. 7). M1 = Primary motor cortex, PFC = Prefrontal cortex. (B) Locations of the seed regions are shown schematically (colored asterisks) on slices of the cerebellum.

Figure 3.

The cerebellum contains at least 4 distinct zones associated with frontal cortex. To illustrate the presence of multiple fronto-cerebellar circuits, maps from distinct frontal seeds are directly compared. Each panel shows the regions being subtracted (left) and the resulting correlation map (right). Maps are at a threshold of z(r) > 0.1. (A) MOT–DLPFC results in preferential correlations with MOT in lobule V in the anterior hemisphere as well as in lobule VIIIB. Preferentially DLPFC-correlated regions include Crus I, Crus II, VIIB, and IX. (B) DLPFC–MPFC further divides the posterior cerebellum: MPFC has greater correlations with Crus I, whereas DLPFC has relatively greater correlations with Crus II (C) MPFC–APFC dissociates in anterior cerebellum betweeen Crus I and lobule VI, respectively. In ventral cerebellum, MPFC preferentially correlates with IX, whereas APFC correlates with VIIIA. (D) APFC–MOT: APFC preferentially correlates with VI, whereas MOT correlates with lobule V in the anterior lobe. APFC continues to correlate with the extent of VI moving ventrally and also appears to correlate with VIIB–VIIIA and Crus II at the ansoparamedian fissure, whereas MOT retains correlations in VIIIB. Numbers refer to the z coordinate plane of the cerebellar slice.

Figure 4.

Fronto-cerebellar circuits in individual subjects. The same comparisons in Figure 3 are computed individually for 3 representative subjects. Results are overlayed on each subject's own anatomical volume. Although the locations of the peak correlations vary somewhat, the overall pattern of functional connectivity is similar to that seen at the group level.

Figure 5.

Raw correlation maps show some bilateral cerebellar connectivity from unilateral cortical seeds. Although subtraction of left and right seeds in a given cortical region highlights the contralateral organization of cerebellar connectivity (see Fig. 1), the raw left and right seeds show present, but weaker, ipsilateral connectivity with the cerebellum. This observation is consistent with the smaller percentage of cerebellar projections that cross back to the ipsilateral hemisphere (see text). Maps are at a threshold of z(r) > 0.1.

Figure 6.

Cerebellar regions are not correlated with primary visual and auditory cortices. Although seeding striate cortex (VIS) and Heschl's gyrus (AUD) produces robust correlations in the cerebral cortex, no connectivity appears to be present in the cerebellum. Correlations with each of the 4 cerebral regions are displayed in successive coronal slices of the cerebellum. Maps are thresholded at z(r) > 0.1. MOT and DLPFC correlations are shown for comparison purposes. The location of the seed regions corresponds to the highest intensity values (white/yellow patches) in the first panel of each column. Numbers correspond to the y coordinate of each coronal slice.

Figure 7.

Neighboring regions of the cerebellum participate in distinct, yet partially overlapping, cerebral networks. (A) Cortical connectivity with bilateral CBM_DLPFC, CBM_MPFC, and CBM_APFC seeds did not show the same strict segregation that was seen in the comparison between CBM_DLPFC and CBM_MOT (Fig. 2). These regions, especially CBM_DLPFC and CBM_MPFC, appear to participate in distributed cortical networks that converge in dorso-, ventro-, and medial PFC, at the posterior midline, and in regions of the lateral parietal and temporal lobes. The CBM_APFC network appears to be segregated from the other 2 networks in the prefrontal cortex, though some convergence was also seen, for example, in BA 47. Borders reflect approximate borders of relevant Brodmann areas encompassing the prefrontal cortex and motor cortex. Hatched regions represent overlap of the CBM_APFC correlation map with the 2 other networks. BA = Brodmann area. (B) Schematic representation of the seed locations (asterisks) on cerebellar slices. CBM_DLPFC coordinates: ±12, −82, −28; CBM_MPFC coordinates: 34, −80, −36 and −32, −76, −34; and CBM_APFC coordinates: ±36, −52, −34.

Figure 8.

A provisional map of human cerebellar topography. All of the data in the present study were combined to provide an estimate of cerebellar topography based on the 4 dissociated regions illustrated in Figure 3. Correlations with the 4 frontal regions are illustrated for descending transverse sections of the cerebellum in the left panel. Each map is based on the averaged (N = 40) z(r) correlation map (threshold = z(r) > 0.1). Hatched regions represent overlap of 2 correlation maps. The z(r) correlation maps are projected onto the cortical surface of the cerebellum in the right panel to illustrate the topographical organization of the fronto-cerebellar connectivity. This map provides a provisional (and certainly incomplete) characterization of the human cerebellum based on connectivity to the frontal cortex. The top projection is a superior view looking down on the rostral and dorsal faces of the cerebellum; the bottom projection shows the view from behind. The middle projection is a rotation between the other 2 (showing the entire dorsal face) to emphasize the relationships among all 4 dissociated cerebellar zones. Note that the majority of the mapped portion of the posterior cerebellum is associated with prefrontal (cognitive) regions of the neocortex. Anatomical labels and major divisions based on the MRI atlas of the human cerebellum (Schmahmann et al. 1999, 2000).

Figure 9.

Fronto-cerebellar circuits dissociate in an independent data set. Spherical seed regions of 2-mm radius were drawn around local maxima in the cerebellar maps generated from Data Set 1. These regions were then carried forward and tested in the independent Data Set 2 to formally quantify the dissociation between the 4 fronto-cerebellar circuits. Each graph depicts the mean z(r) between a given (bilateral) cerebellar region and each of the 4 bilateral frontal target regions. The cerebellar seed regions are depicted in the insets on each graph (coordinates: MOT cerebellar region: ±20, −50, −24; DLPFC cerebellar region: ±12, −80, −24; MPFC cerebellar region: ±22, −86, −40; APFC cerebellar region: ±36, −46, −52). **P < 0.001.