Quantitative Examination of the Bottlenose Dolphin Cerebellum

Authors


Correspondence to: Jacopo Annese, Ph.D.; The Brain Observatory, 3510 Dunhill Street, CA 92121. Fax: 858-534-4064. E-mail: jannese@ucsd.edu

ABSTRACT

Neuroanatomical research into the brain of the bottlenose dolphin (Tursiops truncatus) has revealed striking similarities with the human brain in terms of size and complexity. However, the dolphin brain also contains unique allometric relationships. When compared to the human brain, the dolphin cerebellum is noticeably larger. Upon closer examination, the lobule composition of the cerebellum is distinct between the two species. In this study, we used magnetic resonance imaging to analyze cerebellar anatomy in the bottlenose dolphin and measure the volume of the separate cerebellar lobules in the bottlenose dolphin and human. Lobule identification was assisted by three-dimensional modeling. We find that lobules VI, VIIb, VIII, and IX are the largest lobules of the bottlenose dolphin cerebellum, while the anterior lobe (I–V), crus I, crus II, and the flocculonodular lobe are smaller. Different lobule sizes may have functional implications. Auditory-associated lobules VIIb, VIII, IX are likely large in the bottlenose dolphin due to echolocation abilities. Our study provides quantitative information on cerebellar anatomy that substantiates previous reports based on gross observation and subjective analysis. This study is part of a continuing effort toward providing explicit descriptions of cetacean neuroanatomy to support the interpretation of behavioral studies on cetacean cognition. Anat Rec, 2013. © 2013 Wiley Periodicals, Inc.

Despite the fact that the cerebellum owes its name to the Latin term for “little brain,” it appears as one of the most striking features of the bottlenose dolphin (Tursiops truncatus) brain due to its large size. The cetacean cerebellum has drawn the attention of several neuroanatomists over the last century. Norwegian anatomist Jan Jansen published a definitive article on the subject in 1950, providing a useful template for cetacean cerebellar lobule classification. His analysis derived from observations of twelve embryonic fin whales (Balaenoptera physalus) and one adult bottlenose whale (Hyperoodon rostratus). Olof Larsell (1970), a contemporary of Jansen, was the first to study the cerebellum of the bottlenose dolphin (Tursiops truncatus) specifically. Although Larsell drew heavily from Jansen's classification scheme, his monograph is presently the only available publication that contains detailed information on the anatomy of cerebellar lobules in the bottlenose dolphin.

Though these classic published observations are valuable, they are limited in their usability. Jansen's (1950) and Larsell's (1970) figures, while meticulously drawn, are nearly all based on mid-sagittal sections, which afford clear views of the vermis, but not of the hemispheric lobules. In these studies, insight into the arrangement of hemispheric lobules was gained from surface fissures, which are not always indicative of underlying morphology. Furthermore, Jansen's (1950) and Larsell's (1970) descriptions of lobules were subjective, lacking quantitative comparative measures. Lastly, though both authors commented on the behavioral implications of their anatomical findings, their conclusions were not informed by modern research.

The goal of this study was to derive quantitative measurements from three-dimensional (3D) reconstructions of dolphin and human cerebellar anatomical structures based on magnetic resonance imaging (MRI). These 3D models were based on anatomical delineation of separate lobules conducted on cross-sectional images displayed along three primary orthogonal planes (coronal, sagittal, and horizontal); structural segmentation was the basis for subsequent voxel-based volumetric analysis. There have been previous MRI studies on cetaceans (Marino et al., 2003; Montie et al., 2008; Oelschläger et al., 2008, 2010), including on the bottlenose dolphin (Marino et al., 2000). However, none have focused on the composition of the cerebellum and its lobule boundaries. The purpose of this research is to provide a more detailed account of the bottlenose dolphin cerebellum, a structure that has not been thoroughly analyzed in the context of modern neuroimaging.

We also hypothesize about the implications of the unique dolphin cerebellar structure for its functions, which are unknown. There is evidence that lobule size can be indicative of functional emphasis, and thus it is intriguing to speculate on the significance of the relative lobule sizes (Sultan and Glickstein, 2007). Since no previous studies focus on cetacean cerebellar function, our discussion of lobule function is based on findings in other animals and humans. Though there is a rich history of functional studies on various animals (Burne and Woodward, 1983; Azizi et al., 1985; Horikawa and Suga, 1986), more recent neuroimaging techniques have transformed research in this field, allowing for more precise mapping and new types of functional studies in animals and humans. The cerebellum has long been known to mediate sensorimotor functions, but these recent studies have broadened our understanding of the cerebellum and suggest that the cerebellum may be involved in higher cognitive function as well (Kelly and Strick, 2003; Manni and Petrosini, 2004; Kirschen et al., 2008; Stoodley and Schmahmann, 2009; Balsters et al., 2010; O'Reilly et al., 2010; Buckner et al., 2011).

Functional mapping on the bottlenose dolphin has already begun (Ridgway et al., 2006), and to best interpret this research, it is necessary to have a thorough atlas of the dolphin brain. We hope that our quantitative analysis will aid in future studies on dolphin cognition.

MATERIALS AND METHODS

Data

Tursiops truncatus

Two animals were imaged in the context of this study. Dolphin MK was a male born around 1970 to a bottlenose dolphin population in the Mississippi Sound of the Gulf of Mexico. He was collected in July 1974 for the US Navy Marine Mammal Program and subsequently trained to be a mine-hunting dolphin. He remained with the program until his death, at age 30. During the last year of his life, MK failed to recover from chronic, extensive dermatitis in spite of multiple treatment strategies. After veterinary assessment revealed a serious degradation in the quality of his life, MK was administered euthanasia. Dolphin KL, also male, was collected from the Mississippi Sound in July 1974 for the US Navy Marine Mammal Program. He was trained as a mine-hunting dolphin and remained active in the Program until his death from pneumonia in 2010 at age 40. The behavior and health of both animals were under close monitoring for several years before their death. Neither dolphin showed any evidence of proprioceptive dysfunction or neuromotor deficits that could presumably impact cerebellar structure. Notably, dolphin MK showed a high-frequency hearing impairment (Ridgway and Carder, 1997); however, this is a relatively common condition in older male dolphins and it should not affect our analyses.

MRI scans of the dolphin brains were acquired on a General Electric 1.5 Tesla Excite scanner, using an 8-Channel transfer-receive head coil; in both cases the scans were performed postmortem. The brain of MK was scanned ex situ after perfusion fixation, suspended inside a Plexiglas chamber containing phosphate-buffered formaldehyde solution (scans were acquired at room temperature). The brain of KL was scanned in situ, after perfusion. In both cases, we used two sequences to acquire data with complementary contrast characteristics: a T1-weighted spoiled gradient echo sequence (3D-SPGR; Field of View (FOV): 5122, slice thickness 1 mm; voxel dimensions (1 × 1 × 1 mm) followed by a T2-weighted fast spin echo (2D-FSE; FOV = 10242 slice thickness = 1.4 mm; Repitition Time (TR) = 3,000 ms, Echo Time (TE) = 30 ms; voxel dimension: 0.7, 0.7, 1.4).

Although both dolphin specimens were labeled and lobule volumes quantified, the images displayed in all figures are from dolphin MK.

Homo sapiens

Data from three human cerebella were examined. The scans were acquired in vivo from three neurologically healthy adult volunteers, one male (age 35) and two females (ages 29 and 34). Imaging data were collected with a 3T Siemens Allegra scanner, using a magnetization-prepared rapid-acquisition gradient echo (FOV = 2562, 160 sagittal slices, slice thickness 1 mm, TR = 2,000 ms, TE = 2.1 ms, matrix 192 × 192).

Anatomical labeling

DICOM images were imported into Amira software (Visage Imaging, San Diego, CA). The data was reconstructed and a segmentation editor was used to create delineations on the images. Images were simultaneously viewable in coronal (Fig. 1), axial (Fig. 2), and sagittal (Fig. 3) viewpoints. Figures 4 and 5 display the resulting composite surface images in AMIRA of dolphin MK and a human cerebellum, respectively. The primary references used for labeling the bottlenose dolphin cerebella were Jansen's and Larsell's monographs, mentioned previously (Jansen, 1950; Larsell, 1970), and Haines' atlas of neuroanatomy (Haines, 2007). For the human brains, the main resource used for labeling was the MRI Atlas of the Human Cerebellum by Schmahmann and colleagues (Schmahmann et al., 2000). Table 1 shows the relationship between the different cerebellum lobule and fissure classification systems.

Figure 1.

A–I. Segmentation analysis of the bottlenose dolphin cerebellum (MK) in the coronal plane. The vermal lobules are labeled in red text to distinguish them from the hemispheric lobules, labeled in white. The context of the sections in reference to the rest of the brain from a dorsal perspective is shown in the leftmost column (A,D,G), followed by the unlabeled sections (B,E,H), and finally labeled sections (C,F,I). The red line in the leftmost column indicates a scale of 18 cm. An enlarged view of 1B and C is available in Fig. 8, and 1E and F is seen in Fig. 7. 1A,B,C shows the large lobule IX and the arrangement of the VIII lobules, with VIIIb located more medially than VIIIa. The tuber vermis appears to be left of midline in this section. 1D,E,F displays a middle coronal section, showing the enormous lobule IX, the interior setting of lobule X, and again the off-center VIIaf lobule. 1G,H,I displays a rostral coronal cross-section through the auditory nerve. This section shows that hemispheric and vermal lobules VI are adjacent. Crus I and II are thin, while VIIb and IX are sizeable. Vermal lobule III is visible in this section due to the curvature of the cerebellar vermis.

Figure 2.

A–I. Segmentation analysis of the bottlenose dolphin (MK) in the axial plane. The vermal lobules are labeled in red text to distinguish them from the hemispheric lobules, labeled in white. The context of the sections in reference to the rest of the brain from a posterior perspective is shown in the leftmost column (A,D,G), followed by the unlabeled sections (B,E,H), and finally labeled sections (C,F,I). In the leftmost column, the red scale bar indicates 18 cm. In 2A,B,C, the most ventral slice shown, the large size of lobules VIIb and IX and the interior location of lobule X is apparent. The images in 2D,E,F represent a midlevel section, shows the interior setting of lobule X and affords a clear view of the VIIb, VIII, and IX lobules. Vermal lobules I, II, and III are seen from this section because of the tight curvature of the cerebellar vermis. 2G,H,I is the most dorsal slice, and part of the anterior lobe is visible in this section. This section also displays the continuous nature of the fissures separating lobules VI, and VIIa in both the vermis and hemispheres. The crus lobules appear as thin, single folia.

Figure 3.

A–D. Segmentation analysis of the human (A,B) and bottlenose dolphin MK (C,D) cerebellum in the mid-sagittal plane. The vermal lobules are labeled in red text to distinguish them from the hemispheric lobules, labeled in white. 3C,D displays the primary fissure, which separates the anterior and posterior lobes, and was helpful in beginning the labeling process. Vermal lobule VI appears as three branching folia. The location of the border between VIIaf and VIIat was aided by the ability to analyze multiple sagittal sections. In this section, VIIaf and VIIb appear more internally. Vermal lobule III consists of two thick folia. Part of hemispheric lobule VI is visible in this section.

Figure 4.

A,B. 3D surface reconstruction of a bottlenose dolphin (MK) cerebellum. This figure shows the ventral (A) and dorsal aspects (B) of bottlenose dolphin MK's cerebellum. Hemispheric lobules are labeled in black, while vermal lobules are labeled in red. The relatively small surface representation of the anterior lobe is apparent from a dorsal viewpoint. The enormitiy of lobule IX is visible from a ventral view.

Figure 5.

A,B. 3D surface reconstruction of a human cerebellum. This figure shows the ventral (A) and dorsal (B) view of one of the human cerebella studied. Hemispheric lobules are labeled in black, and vermal lobules are indicated in red.

Table 1. Cerebellar Nomenclature
Schmahmann (1999)Jansen (1950)Jansen abbreviation
Vermis lobules:  
I,IILingula 
IIICentralisl.c.
IV,VCulmencu.
VIDeclivede.
VIIafFolium vermisfo.v.
VIIatTuber vermist.v.
VIIIPyramisp.
IXUvulau.
XNodulusnod.
Hemispheric lobules:  
I,IILingula 
IIICentralisl.c.
IV,VAnterior quadrangular 
VISimplexl.s.
Crus ICrus Icr.I.
Crus IICrus IIcr.II
VIIbParamedianl.pm.
VIIIa, VIIIbParaflocculus dorsalispfl.d.
IXParaflocculus ventralispfl.v.
XFlocculusfl.
Fissures:
I,II–IIIPrecentralf.pc.
III–IVPreculminatef.prc.
IV–VIntraculminatef.ic.
V–VIPrimaryf.pr.
VI–Crus IPosterior superiorf.ps.
Crus I–Crus IIHorizontal 
VIIaf–VIIatHorizontal 
Crus II–VIIbAnsoparamedianf.a.pm.
VIIat–VIIIPrepyramidalf.ppd.
VII–VIII (hemispheres)Parafloccularf.pfl.
VIIIa–VIIIbIntraparafloccularf.i.pfl.
VIII–IXSecundaryf.sec.
IX–XPosterolateralf.pl.

Lobule Labeling of the Bottlenose Dolphin Cerebellum

Anatomical labeling was conducted using 3D visualization and analysis software (Amira), which allowed for simultaneous viewing of coronal, axial, and sagittal images. During the labeling process, we avoided incorporating central cerebellar white matter and nuclei.

We began by labeling the vermis. Primarily, we relied on Jansen's (1950) most late-stage drawings and descriptions of fin whale embryo and adult bottlenose whale cerebella, as well as Larsell's (1970) book chapter on the Cetacean cerebellum, which includes depictions and descriptions of an adult bottlenose dolphin cerebellum.

We first located the primary fissure and lobule VI, which was best visualized from a sagittal viewpoint (Fig. 3C,D). In our specimens, lobule VI is comprised of three large folia, originating from a single medullary ray that splits into three branches (Fig. 3C,D). The primary fissure in deep and distinct, forming the caudal border of the anterior lobe. The layout of the anterior lobe is similar to the depictions of the cetacean species studied by Larsell (1970) and Jansen (1950). Lobule V consists of a large, singular, unbranching medullary ray and folium, of which the rostral border forms the notably straight intraculminate fissure (Fig. 3C,D). Lobule IV consists of two main folia, and is proceeded by the preculminate fissure, marking the edge of lobule III (Fig. 3C,D). Determining the rostral border of lobule III was not as straightforward, as Larsell and Jansen's drawings indicate seemingly contradictory borders. In Larsell's (1970) drawing of a bottlenose dolphin cerebellum, he demarcates lobule III as consisting of one (perhaps faintly branching once near the surface) folium, and subsequently assigning the next two folia as belonging to lobule II. However, our specimens' anatomy appeared much more similar in layout to Jansen's (1950) drawing of a fin whale embryo (Fig. 7, p. 366), in which he designates a second rostral folium as belonging to lobule III. We chose to agree with Jansen (1950), partly because choosing the more caudal fissure as the precentral fissure would create an enormous lobule I/II, consisting of four large folia, which would be a major deviation from both sources. Furthermore, there appears to be a clear fissure between the resulting lobules I and II, further confirming our choice of precentral fissure. Overall, our findings were consistent with written descriptions by both anatomists, with the anterior lobe having a poor hemispheric representation relative to its vermal representation.

After elucidating the boundaries of lobule VI, we established the lobule's posterior boundary as the secondary fissure, which marked the rostral edge of the folium and tuber vermis region. The appearance of thin folia here was reminiscent of one of Jansen's (1950) fin whale embryo (p. 375), signifying the folium and tuber vermis, and nearly identical to one of Larsell's figures of an adult fin whale (Fig. 136B, p. 167). We are confident that we were able to discern both of these vermal lobules, each comprising of a single folium. Utilizing a coronal viewpoint, we followed the path of this horizontal fissure separating the folium and tuber vermis into the hemispheres, which resulted in two thin lobules that were unmistakably the crus lobules, confirming our vermal choices (Figs. 2I and 6). In discerning the vermal lobule VIIb, we first established its boundaries in the hemispheres and traced them back into the vermis, which yielded a small, short section somewhat buried beneath vermal lobule VIII, but clearly continuous with hemispheric lobule VIIb (Fig. 3). Vermal lobule VIIb is much more apparent in the lateral aspects of the vermis than along the sagittal midline.

Figure 6.

A,B. A coronal MRI section of the posterior bottlenose dolphin (MK) cerebellum, both labeled (A) and unlabeled (B). Vermal lobules are labeled in red, hemispheric in white. This section especially showcases the physically continuous relationship between vermal lobules VI, VIIaf, and IX with hemispheric lobules VI, Crus I, and IX, respectively. This section displays the characteristic thinness of Crus II, the interior location of lobule VIIIb and X, and the large size of lobule IX.

Figure 7.

A,B. A labeled (A) and unlabeled (B) coronal MRI section of the bottlenose dolphin (MK) cerebellum at approximately the mid-level of the cerebellum in the coronal orientation. Vermal lobules are labeled in red, hemispheric lobules are labeled in white. This section provides a clear view of hemispheric and vermal lobule X. It is apparent from this view that the bulk of lobule X is interior and likely indiscernible in a superficial examination of the cerebellum, which may have lead previous anatomists to mistakenly devalue its size. As is clearer in the unlabeled MRI section, the actual composition of hemispheric lobule X seems distinct from the other lobules, appearing denser and without long, thin folia. Lobule I, although located farther anterior, is visible in this section. This is due to the angle of the scan and the cyclical curvature of the cerebellum. Lobule VIIat appears laterally alongside vermal lobule VI.

From the anatomists' drawings, it was clear that vermal lobule VIII would likely be distinctive, comprised of several folia that would give a sloping pattern from a sagittal viewpoint. Having established the crus lobules, we were sure of the rostral border of lobule VIII, but unsure of the internal fissure between VIIIa and VIIIb. Utilizing a coronal viewpoint (Fig. 1) was especially helpful in determining the rostral boundary of vermal lobule VIIIb by following its vermal-hemispheric continuum. By following this fissure, it became apparent that hemispheric lobules VIIIa and VIIIb are not distributed evenly as one progresses from medial to lateral sections. While the majority of hemispheric lobule VIIIb is concentrated near the midline, hemispheric lobule VIIIa has the vast bulk of its volume in the lateral portions of the hemispheres (Fig. 1C).

Vermal lobule IX consists of a single, stout medullary ray and folium (Fig. 3C,D). The rostral and caudal fissures remain clear as one moves laterally into the hemispheres. Hemispheric lobule IX is described by Jansen (1950) as occupying “practically the entire ventral surface of the hemispheres” (p. 378), which is consistent with our findings (Fig. 4B).

Because vermal lobule IX's borders are so distinct, it was simple to assign the last remaining cerebellar lobule volume to lobule X. Unlike the other lobules in relation to their neighboring lobules, lobule X is more physically isolated. Its hemispheric component was clearly discernable, especially from a coronal viewpoint (Fig. 7). Furthermore, the actual texture of the lobule seemed distinct from the others, somewhat denser and lacking the frond-like leaflets that border the other folia.

RESULTS

The delineations and surface-based models from the bottlenose dolphin cerebellar MRI data (from MK) and human cerebellar MRI data are shown in Figs. 4 and 5, respectively. Being able to view the cerebellar lobules from multiple planes of sections improved our ability to assess their topological relationships and thus facilitated their classification. Tables 2 and 3 shows the results of the measurements of separate lobule volumes in dolphin MK and KL. According to our data, the vermis occupies approximately 6% (6.6% in MK, 6.1% in KL) of the entire cerebellum of the bottlenose dolphin. Within the vermis, the anterior lobe was the largest, followed in extent by lobules VI, VIIaf, VIIat, VIIb, VIIIa, VIIIb, IX, and X. The lateral hemispheres comprised 93.4% of the cerebellum in MK and 93.9% in KL. The largest lobule was lobule IX, followed by VIIb, VIIIa and VIIIb and VI, Crus I, Crus II, the anterior lobe, and X.

Table 2. Bottlenose Dolphin Lobule Composition
 Entire cerebellum
DolphinPercent vermisPercent hemispheres
MK6.693.4
KL6.193.9
Table 3. Bottlenose Dolphin Lobule Composition
 Percentage of vermis
Vermis lobulesMKKL
I–V39.933.5
VI23.336.7
VIIaf5.15.9
VIIat10.86.1
VIIb2.12.7
VIIIa11.26.3
VIIIb2.64.8
IX4.33.2
X0.70.9
      Percentage of hemispheres
Hemispheric lobules        MKKL
I–V2.02.3
VI7.08.8
Crus_I6.86.1
Crus_II4.45.9
VIIb20.114.5
VIIIa9.112.6
VIIIb9.28.1
IX40.340.5
X1.31.2

We also examined asymmetries between the hemispheric cerebellar lobules of the bottlenose dolphin (Table 5). In both dolphins, Lobule VI showed the greatest degree of asymmetry, with the left side being larger than the right.

The results of the examination of the human cerebellar lobules are shown in Tables 4 and 6, which provides averages for individual lobule volume and the range of values that we collected. Overall, the vermis occupied an average 8% of the cerebellum. The anterior lobe was the largest in the vermis, followed in order of size by lobules VIIIa, VI, IX, VIIat, VIIIb, VIIb, X, and VIIaf.

Table 4. Human Lobule Composition
 Entire cerebellum
SubjectPercent vermisPercent hemispheres
18.092.0
26.693.4
37.992.1
Table 5. Asymmetry in the Bottlenose Dolphin Cerebellum
LobulePercent leftPercent right
(Hemispheres)MKKLMKKL
Anterior (I–V)50535047
VI57614339
Crus I47515349
Crus II48445256
VIIb49465154
VIIIa48535247
VIIIb53524748
IX50505050
X51484952
Table 6. Human Lobule Composition
Vermis lobulesAverage percent of vermisRange (in percent)
I–V38.537.5–40.0
VI11.39.2–12.6
VIIaf2.71.7–3.5
VIIat7.44.6–10.1
VIIb4.94.0–6.7
VIIIa16.215.3–17.7
VIIIb6.24.9–7.7
IX9.26.9–11.3
X3.63.3–3.9
Hemispheric lobulesAverage percent of hemispheresRange (in percent)
I–V9.58.6–10.4
VI16.615.6–17.4
Crus I26.423.0–28.7
Crus II17.714.9–19.3
VIIb8.68.3–9.3
VIIIa7.47.1–7.8
VIIIb5.95.1–6.9
IX7.26.9–7.8
X0.70.6–0.8

The hemispheres of the human cerebellum, therefore, occupied 92% of the entire cerebellum; Crus I was the largest, followed by Crus II, VI, the anterior lobe, VIIb, VIIIa, VIIIb and IX, and X.

DISCUSSION

Cerebellar research was once dominated by cross-species comparisons. The basic anatomical layout of the cerebellum is highly conserved among mammalian species, but the size of the individual lobules can vary greatly (Larsell, 1970; Sultan and Glickstein, 2007). Therefore, researchers assumed that the relative size of each lobule was indicative of functional localization and environmental pressures experienced by each species. Modern research on a variety of mammalian species suggests that the idea of a relationship between lobule size and function still holds some merit (Sultan and Glickstein, 2007). The relationship is strong enough that studies on cerebellum functionality discuss the location of their functional findings in terms of lobule location (O'Reilly et al., 2010; Buckner et al., 2011).

Therefore, as we proceed through the discussion of each lobule, the proposed functions of that lobule will be mentioned as well. Since the patterns of cerebellar connectivity and functionality of the dolphin has yet to be ascertained, we draw from current knowledge of lobule function established in other animals and humans.

Anterior Lobe (I–V)

In the bottlenose dolphin, the anterior lobe (lobules I–V) is much smaller than the posterior (VI–X) lobe (Fig. 4). Interestingly, the anterior vermis lobes comprise 39.9%/33.5% (MK/KL) of the vermis (Fig. 3C,D), but only about 2% of the hemispheres (Fig. 2I). This is consistent with Jansen's (1950, p 372, p 381) findings in both the fin whales and bottlenose whale, as he writes that the anterior lobe is “well developed” in the vermis region, but “rudimentary” in the hemispheres. Larsell (1970, p 167) arrived at a similar conclusion in the bottlenose dolphin, commenting that “the anterior lobe as a whole is smaller in proportion to the entire cerebellum in cetaceans than in any other group of mammals.” In humans, the discrepancy is not as great. From our analysis, the anterior lobe occupies 38.5% of the vermis (Fig. 3A,B), and 9.5% of the hemispheres.

Studies in several species have indicated that the anterior lobe has a strictly sensorimotor role, and contains sensorimotor representations of the body (Manni and Petrosini, 2004; O'Reilly et al., 2010; Buckner, et al., 2011), while the posterior lobe is possibly associated with higher cognitive functions (Kelly and Strick, 2003; Manni and Petrosini, 2004; Kirschen et al., 2008; Stoodley and Schmahmann, 2009). In humans, resting state functional connectivity MRI studies indicate that a secondary sensorimotor representation exists in lobule VIII (O'Reilly et al., 2010; Buckner et al., 2011), one of the largest of the dolphin's lobules. It may be that the functions of these two lobules are redundant and the larger lobule VIII in the dolphin (to be discussed later) simply compensates for the smaller anterior lobe.

Lobule VI

Lobule VI in the bottlenose dolphin is a large component of the vermis (MK/KL: 23.3%/36.7%), but represents a considerably smaller portion of the hemispheres (MK/KL:7.0%/8.8%) (Fig. 4A). Jansen (1950, p 372) found this lobule to be “large in its vermis portion… as well as in its hemisphereal parts” in the fin whales and in the bottlenose whale. On the other hand, Larsell (1970) seemed less impressed by the size of this lobule in the bottlenose dolphin. He directly references Jansen's (1950) findings in the fin whale, stating that the bottlenose dolphin's lobule VI subdivisions appear smaller in comparison.

The wide range of values we found in the two dolphins relative to vermis lobule VI (MK/KL: 23.3%/36.7%) may explain why Jansen (1950) and Larsell (1970) seemed to give opposite descriptions of this lobule's relative size. Our human subjects were more consistent, displaying close ranges for both the vermis (9.2%–12.6%) and hemispheric values (15.6%–17.4%) for this lobule.

Hemispheric lobule VI has well-known sensorimotor functions. In the monkey, Kelly and Strick (2003) found that this lobule projects to the arm area of primary motor cortex. In humans, this lobule is known to contain sensorimotor representation for the tongue, neck, and lip (Manni and Petrosini, 2004). The relatively large head (Jansen, 1950) of the dolphin may explain why the vermis representation is so large in the dolphin.

In a meta-analysis by Stoodley and Schmahmann (2009) of human functional neuroimaging studies, lobule VI was also linked to several types of higher cognitive functions. Lobule VI was likely to be activated during tasks involving working and spatial memory, executive functions, and emotional processing. Some of the proposed memory functions for this lobule overlap with those of the hippocampus, a structure that is unusually small in the bottlenose dolphin (Jacobs et al., 1979). The fact that dolphins excel at memory and spatial tasks while exhibiting reduced hippocampal structures has led some to suggest that these processes occur elsewhere in the brain (Marino, 2004; Montie et al., 2008). It is tempting to speculate that lobule VI may contribute to such functions.

Lobule VIIa (Crus I and II)

The crus lobules are among the smallest in the bottlenose dolphin's posterior lobe, with crus I occupying 6.8%/6.1% of the entire cerebellum and crus II occupying 4.4%/5.9%. Jansen (1950, p 397) found this to be true in both the fin whales and bottlenose whale, commenting that the ansiform lobule is “conspicuously small.” Larsell (1970, p 170) similarly classified these lobules as “narrow” and “relatively small.” He writes that vermis “lobule VII of the porpoise is relatively large,” (1970, p 170) but was unable to differentiate between vermis lobules VIIat and VIIaf. However, using our 3D approach, we were able to locate the division (Fig. 3C,D). Certain coronal (Figs. 6 and 8) and axial (Fig. 9) viewpoints show that this division may have been difficult to locate from a gross anatomical perspective because of the asymmetrical arrangement of vermal lobule VII. A single midline section could easily miss the bulk of one or both of these lobules, making its border challenging to distinguish.

Figure 8.

A,B. A labeled (A) and unlabeled (B) coronal MRI section of the posterior bottlenose dolphin (MK) cerebellum slightly posterior to Fig. 6. Vermal lobules are labeled in red, hemispheric in white. This section displays the continuous nature of the vermal/hemispheric fissures between lobules VIIat/Crus II and VIIIb/VIIIb. In this section, vermal lobules VIIat and VIIIa are not symmetrical along the midline. A small part of vermal lobule IX is apparent in this section, bordering its hemispheric aspect, which is notably large. This viewpoint of the intraparafloccular fissure is useful in determining the continuation of the vermal fissure separating VIIIa and VIIIb. The bulk of hemispheric lobule VIIIa is located more laterally, while lobule VIIIb is concentrated medially.

Figure 9.

A,B. A labeled (A) and unlabeled (B) axial MRI section of the bottlenose dolphin (MK) cerebellum at approximately the mid-level of the cerebellum in the axial orientation. Vermal lobules are labeled in red, and hemispheric lobules in white. This section illustrates the physical continuity of lobule VIIb into its vermal sections. Vermal lobule VIIb appears to taper significantly toward the midline, and is much more voluminous near the lateral vermis. This section also showcases the thinness of the crus lobules and the somewhat asymmetrical representation of their vermal components, which may explain the difficultly previous anatomists experienced in locating the Horizontal fissure when examining single mid-sagittal sections.

In contrast with dolphins, the human hemispheric crus I and crus II are the largest lobules, occupying an average 26.4% and 17.7% of the cerebellar hemispheres, respectively. Jansen (1950) speculated that the crus lobules were involved in limb functions, which explains why they would be relatively small in the dolphin. Experiments in cats by Snider and Stowell (1944) found that both crus lobules were activated during stimulation of the forepaws. However, recent evidence has shown that, in some species, the crus lobules are likely involved in higher cognitive functions. Using neuronal tracers in monkeys, Kelly and Strick (2003) discovered that crus II both projects to and receives information from area 46 in prefrontal cortex. Resting state functional connectivity experiments in humans by Buckner et al. (2011) and O'Reilly et al., (2010) agree with those results, confirming that activity in crus I and crus II is correlated with activity in association cortices, such as the prefrontal cortical areas and the posterior parietal cortex. Balsters and Ramnani (2008) found this lobule was activated when human subjects were interpreting symbolic cues, a type of higher cognitive function associated with prefrontal cortex. The large size of the crus lobules in the human cerebellum supposedly reflects the characteristic enlargement of the prefrontal cortex (Balsters et al., 2010) in our species.

Likewise, the lack of a homologous prefrontal lobe in the dolphin may explain why the crus lobules are relatively small. It is noteworthy, however, that although this particular structure is absent or greatly reduced in the dolphin, many of the functions attributed to this region are not (Marino, 2004; Herman, 2006). Reiss and Marino (2001), for example, found that dolphins are capable of self-awareness, an ability previously thought to be limited only to certain primates boasting the necessary prefrontal cortical structures. Similar to the conundrum of displaying an excellent memory despite a small hippocampus, it is likely that some of these prefrontal functions are present in the dolphin brain, but simply located elsewhere.

Lobule VIIb

Lobule VIIb is large (20.1%/14.5% of the hemispheres) in the bottlenose dolphin (Fig. 1C, F and 7) as compared to the human (only 8.6% of the hemispheres). Our findings agree with those of Jansen (1950, p 397), who describes this lobule in both his specimens to be “good-sized.” Referring to the bottlenose whale, Jansen (1950, p 381) writes that “the lobulus paramedianus is large, exceeding the ansiform lobule in size.” Larsell (1970, p 172) arrived at a similar conclusion, stating that “the paramedian lobule is the second largest division of the cerebellum and forms a large portion of the hemisphere.”

The somewhat sparse studies of function in this lobule indicate that it likely has a sensorimotor role. In the cat, Snider and Stowell (1944) found evidence for sensorimotor connections with the fore and hind limbs as well as activation in response to auditory cues. In Kelly and Strick's 2003 labeling experiments in monkeys, this lobule was found to have connections with the arm area of primary motor cortex. It would be surprising if the role of lobule VIIb in the dolphin were related to limb functions since these structures are relatively rudimentary in the dolphin. Given the lobule's large size, one would be more inclined to speculate that its function is auditory-associated, as evidenced in the cat.

Lobule VIIIa and b

Lobules VIIIa and b are collectively the third largest component of the bottlenose dolphin cerebellum (9.1%/12.6% for VIIIa and 9.2%/8.1% for VIIIb), and are slightly larger than in humans subjects (7.4% for VIIIa and 5.9% for VIIIb). Both Jansen (1950) and Larsell (1970, p 171) agree that this lobule is hypertrophied in cetacean species, though Larsell claims it is “especially pronounced” in the bottlenose dolphin.

Lobule VIII collectively is thought to be concerned with sensorimotor functions, especially audition. As mentioned earlier, a sensorimotor map of the body exists in this lobule (Manni and Petrosini, 2004; O'Reilly et al., 2010). In the cat, Snider and Stowell (1944) found this lobule was activated by auditory cues. A recent study by Kirschen, et al. (2008) found that children who have undergone cerebellar tumor resection in this lobule often display reduced digit span for auditory memory. Larsell (1970) mentions that, in the echolocating bat, this is the only lobule that appears highly developed. Interestingly, another study done on bats by Horikawa and Suga (1986) indicated that this hemispheric lobule (along with hemispheric lobule IX) was likely involved in stabilization of emitted echolocation pulses.

This auditory-associated lobule would be expected to be large in the bottlenose dolphin since this species relies heavily on the auditory sense. In directing echolocation trains and responding to this information, the dolphin must adjust its body quickly and precisely according to echolocation signals. This lobule may provide a location for such physical and auditory integration to occur.

Lobule IX

Lobule IX is the largest lobule in the dolphin cerebellum. By our estimate, it occupies 40.3%/40.5% of the cerebellar hemispheres. In humans, lobule IX accounts for only 7.2% of the hemispheres. Jansen (1950, p 372) was similarly impressed with the size of this lobule in both of his cetacean specimens, commenting (on the fin whales) that the “paraflocculus ventralis is by far the largest subdivision of the cerebellum, comprising the ventral and the caudal surface of the hemisphere,” which agrees with our delineations (Fig. 4). Larsell (1970, p 171) claimed that the bottlenose dolphin's lobule IX is even “larger than in the fin whale.”

Jansen (1950) postulated that this lobule was likely related to auditory function. Evidence has since supported that theory, with the addition that visual functions may also be involved. In 1983, Burne and Woodward found evidence that the visual cortices of the rat project to the ventral paraflocculus via the pons. Along with the cerebellum, the pons is known to be characteristically large in the bottlenose dolphin, as well as in cetaceans in general (Oelschläger et al., 2008). In 1985, Azizi et al.'s research in rats revealed that the paraflocculus also receives input from cerebral auditory cortex through the pons. The researchers found that stimulation of the auditory cortex or the inferior colliculus (another structure typically hypertrophied in cetaceans) elicited responses in the rat paraflocculus. Soon after, Horikawa and Suga (1986) found that lobule IX appears to be involved in the production of echolocation pulses in bats. It is also worth noting that lobule IX physically envelops part of the eighth cranial nerve, which brings auditory information into the brain.

Although the functions of lobule IX have never been studied in the living dolphin, studies in other animals indicate that it is often visual and auditory associated. Because of the dolphin's reliance on echolocation, our finding that lobule IX is the largest lobule in the bottlenose dolphin cerebellum implies that its functions may be similar. Our evidence supports the theory that the large size of this lobule could reflect the dolphin's ability to survey and navigate its environment using both echolocation and vision.

Lobule X

In both his specimens, Jansen (1950, p 397) classifies the flocculonodular lobe as rudimentary. Larsell (1970, p 174) simply states that lobule X is “small in cetaceans.” In our bottlenose dolphins, we also found the flocculonodular complex to be a relatively small component of the cerebellum, occupying 1.3%/1.2% of the cerebellar hemispheres.

However, one wonders exactly which other species Larsell (1970) and Jansen (1950) had in mind when classifying the cetacean cerebellar lobule X as “small.” In our human cerebella, the volume of the lobule X complex was about 0.7% of the hemispheres. One possible reason for Jansen's (1950) and Larsell's (1970) observations is that the main bulk of hemispheric lobule X in the bottlenose dolphin is concealed by the enormous lobule IX (Figs. 2C, 6, and 7). It is only possible to see the full volume of lobule X by analyzing non-medial cross-sections of the cerebellum (Fig. 7). As mentioned earlier, Jansen (1950) and Larsell (1970) relied heavily on fissures visible from the exterior when making conclusions about hemispheric lobules. Since lobule IX is so hypertrophied, only a misleadingly small part of lobule X appears on the surface.

Nonetheless, although the hemispheric representation of X is similar, the vermis representation is much greater in our humans (3.6% vs. 0.7%/0.9%). Across species, the flocculonodular lobe is known to receive input from the vestibular system, and plays an important role in maintaining balance and directing eye movement. The vestibular system is significantly reduced in dolphins (Oelschläger et al. 2008). One important function of the vestibular system is to sense the movement of the head and to modify the perception of sound accordingly. Oelschläger et al., (2008) theorizes that the small vestibular system may be a result of decreased overall mobility of the head in dolphins. Jansen (1950) proposed that the reduced size might be due to decreased demands on balance in an aquatic environment.

Asymmetry

The lobules of the posterior vermis appear asymmetrical in both dolphin specimens (Fig. 4). This observation is likely not a result of the fixation process since dolphin MK's brain was perfused before removal from the skull, and dolphin KL's brain was scanned in vivo. Additionally, both Jansen (1950) and Larsell (1970) noted this vermal asymmetry in their cetacean species. We wondered whether the asymmetry seen in the vermis could be indicative of volumetric asymmetry in the hemispheres, a fact that was not mentioned in either publication. Hence, we measured the relative volumes of the right and left lobules in the dolphin cerebellum (Table 5). Nearly all lobules were evenly distributed (50% ± 4%), with the exception of lobule VI, which was considerably larger on the left side (57% for MK and 61% for KL) than the right (43% and 39%, respectively) in both dolphins. Since most connections are crossed between the cerebellum and cerebral cortex, it is interesting that Ridgway and Brownson (1984) found a slightly larger surface area of the right cerebral cortex in Tursiops truncatus.

The cerebral hemispheres of the dolphin brain are believed to be quite independent. The corpus callosum is thin in this species (Ridgway and Tarpley, 1994). Dolphins are known to undergo unihemispheric slow-wave sleep, and are able to remain vigilant even with only one hemisphere awake (Ridgway et al., 2006). Positron emission tomography (PET) scans during induced unihemispheric slow-wave sleep reveal that there is also lateralization in cerebellar activity (Ridgway et al., 2006). However, the fact that dolphins can remain vigilant even with one hemisphere asleep implies that there is no extreme lateralization of function that would cause severe impairment.

Conclusions will remain speculative until lobule functions in the dolphin are better understood. Since cerebellar connections to the cerebral cortex are mostly crossed, a large left-sided lobule VI may imply a specialized right-sided structure in the cortex. As mentioned previously, Ridgway and Brownson (1984) did observe a slightly larger surface area in the right cerebral hemisphere. Though no abnormalities have been noted during vigilance testing to imply lateralization of particular functions, perhaps future research will reveal some.

CONCLUSIONS

There are limitations to this study that we would like to acknowledge. First and foremost is the fact that only two brains were used for the bottlenose dolphin analyses. However, postmortem anatomical atlases are generally based of one or very few representative specimens and there is no reason to suppose that Dolphin MK or KL had abnormal brains or were outliers for their species. Furthermore, not only were the cerebellar composition of dolphin KL and MK similar to one another, but their features were also consistent with the distinctively cetacean characteristics described by previous studies (Jansen, 1950; Larsell, 1970). Second, the MRI scans that we examined were qualitatively different. Though both dolphins were scanned postmortem, MK was scanned ex situ after perfusion fixation, and KL was scanned on the day of his death in situ. Additionally, we used scans of fixed tissue in the case of dolphin subjects and fresh (in vivo) scans of human volunteers. Concerns over potential discrepancies between datasets are reduced by the fact that our comparisons are based on relative measurements within each specimen. It is unlikely that the process of fixation would affect separate lobules differently; rather, it is more probable that processing artifacts created changes in volume and contrast that affected the cerebellum as a whole. Furthermore, a study by Montie, et al. (2010) on the effects of fixation on brain structure volume in sea lions found little discernible difference in volumes of white or gray matter in the cerebrum between scans of live sea lions compared to freshly deceased brains and brains that had undergone fixation.

Another limitation in the research is the human error involved in the segmentation analysis. Though we sought to avoid including white matter in our lobule labeling, it is possible that small inclusions were made due to human error and visualization limitations. Also, a small degree of inaccuracy may have affected our segmentation of structures based on MRI. Partial volume effects due to the resolution of the images could have contributed to some error in tissue classification (i.e., white versus grey matter selection). The possibility of examining the volumetric data in 3D lessens this concern to some degree.

Finally, lobule functions are speculative, so the discussion of the relationship between structure and function in the dolphin brain is admittedly tentative. Most studies concerning the relationship between the cerebellum and higher cognitive functions have derived from studies on human subjects and it is presently unclear how these findings are applicable to dolphins. Specialized imaging techniques, such as Diffusion Tensor Imaging (DTI; Mori and Zhang, 2006) will be particularly useful to elucidate patterns of connectivity in this species and other marine mammals.

Overall, our classification of the bottlenose dolphin cerebellum contains the same features described in Jansen's (1950) fin and bottlenose whales and Larsell's (1970) bottlenose dolphin. The anterior lobe is smaller than the posterior, more so in the hemispheres than in the vermis. Lobule IX is largest and dominates the ventral aspect of the cerebellum. Lobules VIIb and VIII are also sizeable. The crus lobules are relatively small, especially compared to those in the human cerebella. Overall, the similarity between the three cetacean species is useful, and implies that this cerebellar atlas may be helpful when studying various cetacean species.

Minor differences, however, do exist. For example, we found conflicting evidence concerning lobule X, which both Jansen (1950) and Larsell (1970) regarded as small. When compared to the relative size of lobule X in the human, the dolphin version is at least comparable, if not actually larger. Furthermore, in regards to lobule VI, we could not agree with either Larsell (1970) or Jansen (1950), since MK and KL gave different relative volumes, indicating that perhaps the size of this particular lobule can vary in individual bottlenose dolphins. Regardless, our method of quantifying the lobule values reduces the ambiguity created by using purely qualitative terminology.

We were also able to make observations that were beyond the scope of the work of Jansen (1950) and Larsell (1970). Our approach allowed for the visualization of the interior of the lobules, which is how we established that lobule X was not as diminutive as Larsell (1970) and Jansen (1950) described. Additionally, we were able to locate both lobule VIIat and VIIaf in the vermis, which Larsell was unable to do. We also more thoroughly analyzed the cerebellar lobule asymmetry apparent from the cerebellar surface. Though Jansen (1950) and Larsell (1970) commented on the obvious asymmetry in the posterior vermis, our method allowed for an investigation of the subtler asymmetry in the hemispheric lobules. Though most lobule volumes were approximately evenly split across the vermis, we found that lobule VI was larger on the left side in both bottlenose dolphins.

After reviewing recent research into cerebellar function, it is likely that the dolphin's large cerebellum is due to the species' reliance on auditory information. Indeed, the largest lobules in the dolphin, IX, VIII and VIIb, have all been linked to auditory function in other species. However, as previously stated, it may be premature to assert a functional role without knowledge of connectivity patterns. Research into cerebellar function and dolphin cerebellar connectivity, using novel techniques such as DTI, will likely impact our understanding of the bottlenose dolphin brain and of the cerebellum in particular. It is our hope that these studies will benefit from the morphometric data provided by this communication.

ACKNOWLEDGEMENTS

This work is part of a comprehensive effort towards creating an inter-operable 3D atlas of the entire dolphin brain. Postmortem scans of the dolphin brains were conducted at the Radiology Imaging Laboratory (Department of Radiology, UC San Diego) under The Brain Observatory's neuroimaging program. The authors thank Miss Chris Ha, research associate at The Brain Observatory for her technical assistance and critical evaluation of the manuscript. The Institutional Animal Care and Use Committee of the US Navy Marine Mammal Program provided oversight of all research with the two dolphins. Thanks to Russell A. Poldrack, Ph.D., Professor of Psychology and Neurobiology, Director of the Imaging Research Center, U of Texas, Austin for providing MRI images of the human brains. The human study was approved by the Institutional Review Board at UCLA, and informed consent was obtained from all volunteers prior to participation.

Ancillary