disease of aquatic organisms 91:243

DISEASES OF AQUATIC ORGANISMSDis Aquat Org

Vol. 91: 243–256, 2010doi: 10.3354/dao02259

Published September 17

INTRODUCTION

Magnetic resonance imaging (MRI) is emerging as auseful tool for studying the neuroanatomy of marinemammals. MRI-based atlases have been completed fornumerous species, including the beluga whale Delphi-napterus leucas (Marino et al. 2001a), common dolphinDelphinus delphis (Marino et al. 2001b), bottlenosedolphin Tursiops truncatus (Marino et al. 2001c), har-bor porpoise Phocoena phocoena (Marino et al. 2003a),

dwarf sperm whale Kogia simus (Marino et al. 2003b),spinner dolphin Stenella longirostris orientalis (Marinoet al. 2004a) and killer whale Orcinus orca (Marino etal. 2004b). These atlases were created from MRI scansof brains that were removed from the skull and forma-lin fixed. To avoid dissection and fixation artifacts,neuroanatomy studies of the Atlantic white-sided dol-phin Lagenorhynchus acutus were completed fromMRI scans of fresh, postmortem brains intact within the

© Inter-Research 2010 · www.int-res.com*Email: [email protected]

Magnetic resonance imaging quality and volumesof brain structures from live and postmortemimaging of California sea lions with clinical

signs of domoic acid toxicosis

Eric W. Montie1, 2,*, Elizabeth Wheeler2, Nicola Pussini2, Thomas W. K. Battey3, Jerome Barakos4, Sophie Dennison2, Kathleen Colegrove5, Frances Gulland2

1College of Marine Science, University of South Florida, Florida 33701, USA2Veterinary Science Department, The Marine Mammal Center, Sausalito, California 94965, USA

3Eckerd College, Galbraith Marine Science Center, St. Petersburg, Florida 33711, USA4California Pacific Medical Center, University of California, San Francisco, California 94143, USA

5Zoological Pathology Program, College of Veterinary Medicine at Urbana-Champaign, Maywood, Illinois 60153, USA

ABSTRACT: Our goal in this study was to compare magnetic resonance images and volumes of brainstructures obtained alive versus postmortem of California sea lions Zalophus californianus exhibitingclinical signs of domoic acid (DA) toxicosis and those exhibiting normal behavior. Proton density-(PD) and T2-weighted images of postmortem-intact brains, up to 48 h after death, provided similarquality to images acquired from live sea lions. Volumes of gray matter (GM) and white matter (WM)of the cerebral hemispheres were similar to volumes calculated from images acquired when the sealions were alive. However, cerebrospinal fluid (CSF) volumes decreased due to leakage. Hippocam-pal volumes from postmortem-intact images were useful for diagnosing unilateral and bilateral atro-phy, consequences of DA toxicosis. These volumes were similar to the volumes in the live sea lionstudies, up to 48 h postmortem. Imaging formalin-fixed brains provided some information on brainstructure; however, images of the hippocampus and surrounding structures were of poorer qualitycompared to the images acquired alive and postmortem-intact. Despite these issues, volumes of cere-bral GM and WM, as well as the hippocampus, were similar to volumes calculated from images of livesea lions and sufficient to diagnose hippocampal atrophy. Thus, postmortem MRI scanning (eitherintact or formalin-fixed) with volumetric analysis can be used to investigate the acute, chronic andpossible developmental effects of DA on the brain of California sea lions.

KEY WORDS: Domoic acid · California sea lion · Magnetic resonance imaging · MRI · Brain ·Hippocampus · Marine mammal

Resale or republication not permitted without written consent of the publisher

OPENPEN ACCESSCCESS

Dis Aquat Org 91: 243–256, 2010

skull, with the head still attached to the body (Montieet al. 2007, 2008). More recently, a neuroanatomicalMRI-based atlas was completed from magnetic reso-nance (MR) images of a live California sea lion Zalo-phus californianus (Montie et al. 2009).

Regardless of how the brain is imaged (i.e. live, freshand intact within the skull, removed and formalinfixed), MRI has its advantages over traditional histo-logical methods. MRI allows thin virtual sections(~2 mm in thickness) of the entire brain to be acquiredefficiently and quickly, where gross and histologicalsectioning is not practical for viewing the entire brain.Thus, MRI can be used as a screening tool to identifylesions for further histological analysis. In addition,MRI coupled with image analysis can accurately deter-mine the volumes of brain structures (a techniquecalled volumetric neuroimaging), while traditional dis-section and photography introduces error in perform-ing quantitative measurements. Three-dimensional(3D) models of brain structures constructed from MRIscans can provide a valuable tool to examine spatialrelationships among brain structures. In fact, volumet-ric neuroimaging and 3D modeling have illustratedtheir powerful utility in studying the brains of theAtlantic white-sided dolphin and California sea lion(Montie et al. 2008, 2009).

Traditionally, MRI has been used as a tool in diag-nosing causes of neurological disease. In the case ofmarine mammals, there is a risk of neurological disor-ders associated with exposure to: (1) anthropogenicchemicals, such as polychlorinated biphenyls (PCBs)(Zoeller et al. 2002, Sharlin et al. 2006) and polybromi-nated diphenyl ethers (PBDEs) (Costa & Giordano2007), that affect neurodevelopment; (2) terrestrialpathogens that can cause brain lesions (Conrad et al.2005); (3) noise pollution that may disrupt normal divebehavior and potentially cause gas-bubble lesions inthe brain (Jepson et al. 2003); and (4) marine neuro-toxins associated with harmful algal blooms that cancause necrosis in brain structures such as the hippo-campus (Scholin et al. 2000, Silvagni et al. 2005). MRIand volumetric neuroimaging would be very usefultechniques to assess the effects of these stressors onthe neurological health of marine mammals.

A rising health concern for California sea lions is neu-rological disease associated with exposure to domoicacid (DA), a marine neurotoxin produced by diatomsbelonging to the genus Pseudo-nitzschia (Scholin et al.2000, Silvagni et al. 2005, Goldstein et al. 2008, Rams-dell & Zabka 2008). In 1998, more than 400 sea lionswere exposed to DA through contaminated prey(Scholin et al. 2000). Sea lions that died acutely andcontained detectable levels of DA in blood and urineexhibited lesions in the hippocampus (Scholin et al.2000). These lesions were characterized by neuronal

necrosis in the hippocampal formation, leading to atro-phy of this brain structure (Silvagni et al. 2005). Evi-dence also suggests that chronic exposure to sublethallevels of DA can lead to varying degrees of hippocam-pal atrophy, both unilateral and bilateral, in Californiasea lions (Goldstein et al. 2008). Now, there is concernthat low levels of DA exposure in the developing fetusand neonate, as well as co-exposure to anthropogenicpollutants such as dichlorodiphenyltrichloroethane(DDTs), may cause subtle changes in the brain that mayresult in long-term cognitive impairment (Ramsdell &Zabka 2008, Tiedeken & Ramsdell 2010). Volumetricneuroimaging would be a useful method to determinesubtle changes in morphology of the hippocampus dueto varying DA exposure levels and scenarios.

In some cases, due to logistical and ethical concerns,it is advantageous to image the brain postmortem,either intact within the skull, with the head stillattached to the body (i.e. postmortem-intact), or re-moved and formalin fixed (i.e. postmortem-fixed).Postmortem changes in the brain may affect MRI qual-ity, introduce artifacts and affect volumes, makingdiagnosis of DA toxicosis difficult. Our goal in thisstudy was to compare MR images and volumes of brainstructures obtained from live and postmortem MRIscans of non-DA and DA intoxicated California sealions. Specifically, the objectives were to: (1) compareMR images and brain structure volumes obtained fromMRI scans completed on live sea lions, then on brainsat 2, 24, and 48 h postmortem-intact; (2) compare MRimages and brain structure volumes obtained fromMRI scans completed on live sea lions versus imagingcompleted on brains that were postmortem-fixed; and(3) determine the usefulness of postmortem imaging indiagnosing DA toxicity.

MATERIALS AND METHODS

Source and processing of samples. California sealions used in this study stranded live along the Cali-fornia coast from Marin to San Luis Obispo countiesbetween 2007 and 2008 (Table 1). Sea lions were takento The Marine Mammal Center, Sausalito, for clinicalassessment, treatment and rehabilitation. Sex wasdetermined based on genital morphology, while ageclass was based on body length, tooth size and stage ofsagittal crest development (Greig et al. 2005). Bloodsamples were collected for serology, serum biochem-istry and hematology. Domoic acid intoxication wascharacterized by clinical signs that included ataxia,head weaving or seizures. Symptoms also includedmarked lethargy and inappetance, vomiting, musculartwitching and central blindness blepharospasm (Gold-stein et al. 2008). Serum, urine and feces samples were

244

Montie et al.: MRI of sea lions with domoic acid toxicosis

analyzed for DA by direct competitive DA ELISA(Biosense Laboratories).

Because prognosis of all sea lions in this study waspoor, they were euthanized, which allowed for thecomparisons of MR images and volumes of brain struc-tures obtained and derived from live and postmortemscanning. The first data set comprised MRI scans of livesea lions performed on animals nos. CSL 7552 (named‘Gratitude’), CSL 7775 (‘Kirina’) and CSL 7999 (‘Barlich’).The brains of these sea lions were then scanned 2, 24 and48 h postmortem, with the brain intact within the skulland the head attached to the body (i.e. postmortem-intact). In between postmortem-scan time points, speci-mens were temporarily stored at 4°C. The second dataset comprised MRI scans of live sea lions performed onCSL 7968 (‘Fairbanks’), CSL 7964 (‘Tiki’), CSL 7975 (‘Ru-pert’), CSL 7949 (‘Tintoretto’), CSL 7999 (‘Barlich’) andCSL 7552 (‘Gratitude’). At necropsy, the brains of thesesea lions were removed and immersed in 10% neutralbuffered formalin. The elapsed time from death to im-mersion of brain in formalin varied between specimens(Table 2). Specimens were stored at 4°C until the brainwas removed and fixed. At the time of MRI, the fixedbrain was removed from the formalin container, dried,placed in a Ziploc bag and scanned (i.e. postmortem-fixed). The elapsed time from immersion of the brain informalin to brain scanning varied between specimens(Table 2). Following postmortem MRI, brains were sec-tioned for histology in the oblique plane perpendicular tothe long axis of the sylvian fissure and temporal lobe.Tissues were processed routinely, embedded in paraffin,sectioned at 5 µm and stained with hematoxylin andeosin (H&E). All studies were conducted in accordancewith institutional, national and international guidelinesconcerning the use of animals in research.

Magnetic resonance data acquisition. Radiographsof the head and chest of each individual were taken torule out the presence of any metal that would result inimage artifacts and/or injure the sea lion. Live sea lionsanesthetized with isoflurane, intact specimens, andfixed brains were imaged with a 1.5-T Siemens Mag-netom Symphony scanner (Siemens) equipped with

245

Acc

essi

onS

ea l

ion

Ad

mis

sion

Str

and

loc

atio

nS

exA

ge

Len

gth

Bra

inE

tiol

ogy

DA

in

MR

Is c

omp

lete

dn

o.n

ame

dat

e d

/mo/

yr(c

m)

wei

gh

t (g

)fe

ces

(ng

g–

1 )

CS

L 7

552

Gra

titu

de

21/1

1/20

07A

lam

eda

Cou

nty

, Alb

any

FS

ub

adu

lt13

735

4.7

Dom

oic

acid

4.8

live

, 2, 2

4, 4

8, f

orm

alin

CS

L 7

775

Kir

ina

23/0

7/20

08S

an L

uis

Ob

isp

o C

oun

ty,

FY

earl

ing

117

306

Ost

eom

yeli

tis

<4.

0li

ve, 2

, 24,

48

San

Sim

eon

hin

d f

lip

per

CS

L 7

949

Tin

tore

tto

04/1

0/20

08M

arin

Cou

nty

, Poi

nt

Rey

es

FA

du

lt13

9n

aD

omoi

c ac

id<

4.0

live

, for

mal

inN

atio

nal

Sea

shor

eC

SL

796

4T

iki

12/1

0/20

08S

an L

uis

Ob

isp

o C

oun

ty,

FY

earl

ing

116

348.

3D

omoi

c ac

id<

4.0

live

, for

mal

inP

ism

o B

each

CS

L 7

968

Fai

rban

ks

14/1

0/20

08S

anta

Cru

z C

oun

ty, L

aSel

vaM

Juve

nil

e14

835

5.3

Dom

oic

acid

<4.

0li

ve, f

orm

alin

CS

L 7

975

Ru

per

t17

/10/

2008

Ala

med

a C

oun

ty, B

erk

eley

MJu

ven

ile

147

na

Lep

tosp

iros

is<

4.0

live

, for

mal

inC

SL

799

9B

arli

ch05

/12/

2008

Mar

in C

oun

ty, D

illo

n B

each

FA

du

lt15

536

1D

omoi

c ac

id33

.6li

ve, 2

, 24,

48,

for

mal

in

Tab

le 1

. Z

alop

hu

s ca

lifo

rnia

nu

s.S

tran

din

g a

nd

lif

e h

isto

ry i

nfo

rmat

ion

of

Cal

ifor

nia

sea

lio

ns.

Mag

net

ic r

eson

ance

im

agin

g (

MR

I) o

f th

e b

rain

s of

th

ese

sea

lion

s w

asco

mp

lete

d li

ve a

nd

2, 2

4 an

d 4

8 h

pos

tmor

tem

, wit

h t

he

bra

in in

tact

wit

hin

th

e sk

ull

an

d t

he

hea

d a

ttac

hed

to

the

bod

y an

d/o

r re

mov

ed a

nd

for

mal

in f

ixed

. Dat

es a

re g

iven

as

dd

/mm

/yea

r. D

A: d

omoi

c ac

id; n

a: n

ot a

vail

able

Sea lion Elapsed time: Elapsed time:name death to fixation (h) fixation to MRI (d)

Gratitude 51 518Tintoretto 23 210Tiki 3.5 199Fairbanks 2 196Rupert 4 191Barlich 48 141

Table 2. Zalophus californianus. Time to fixation and mag-netic resonance imaging (MRI) of California sea lion brains

Dis Aquat Org 91: 243–256, 2010

a CP Extremity Coil. After the localizer scan, T1-weighted images in the sagittal plane were acquiredusing a spoiled gradient echo (FLASH) sequence withthe following parameters: TR = 22 ms, TE = 10 ms,FOV = 200 × 200 mm, slice thickness = 1 mm, and voxelsize = 0.3 × 0.3 × 1 mm. Two-dimensional proton den-sity- (PD) and T2-weighted images in the transverseplane were acquired using a turbo spin-echo (TSE)sequence with the following parameters: TR = 3650 ms,TE = 14/98 ms for PD and T2, respectively, slice thick-ness = 2.5 mm, FOV = 150 × 150 mm, and voxel size =0.3 × 0.3 × 2.5 mm. Additionally, 2-dimensional PD-and T2-weighted images in the oblique plane (i.e. per-pendicular to the long axis of the sylvian fissure andtemporal lobe) were acquired using a TSE sequencewith the following parameters: TR = 5470 ms, TE =14/98 ms for PD and T2, respectively, slice thickness =2.5 mm, FOV = 160 × 160 mm, and voxel size = 0.3 × 0.3× 2.5 mm. The oblique orientation was selected tooptimize viewing of the hippocampus, as previouslydescribed (Goldstein et al. 2008, Montie et al. 2009).

Anatomic identification. Anatomical structures wereidentified using the MRI-based brain atlas of the Cali-fornia sea lion (Montie et al. 2009), the brain atlasof the domestic dog Canis familiaris (beagle) (Dua-Sharma et al. 1970) and that of the human (Nolte &Angevine 2000). Brain structures were labeled usingnomenclature adopted from the English translation ofNomina Anatomica Veterinaria (ICVGAN 2005). MRimages were created using eFilm Lite 2.1.2 (MergeHealthcare) from the Digital Imaging and Communica-tion in Medicine (DICOM) images saved during theFLASH and TSE sequences. Left and right orientationof the images followed the traditional radiology format,where the left side of the brain appears on the rightside of the image.

Volume analysis of brain structures. Initial evalua-tion of MR images was completed at the MRI unit.Post-processing, segmentation (i.e. assigning pixels toparticular structures), 3D reconstructions and volumeanalysis were performed using the software programAMIRA 4.1.1 (Mercury Computer Systems). The samemethods as those described in the MRI-based brainatlas of a live California sea lion (Montie et al. 2009)were used to segment brain structures, create 3D re-constructions and perform volume analysis.

For live versus 2, 24 and 48 h postmortem-intactcomparisons, 3D reconstructions and volumes of brainstructures were determined for (1) whole brain, (2)cerebrospinal fluid (CSF) of the total, left and rightbrain, (3) CSF of the total, left and right cerebral ven-tricles, (4) GM and WM of the entire brain, (5) GM andWM of the total, left and right cerebral hemispheres,and (6) GM and WM of the cerebellum including thebrainstem. Volumes of the left and right hippocampus

and associated structures were determined for (1) lat-eral ventricle (ventral horn), (2) hippocampal sulcus(which included CSF of sulcus and CSF of transversefissure of bichat), (3) hippocampus (which included thesubiculum, cornu ammonis, dentate gyrus, fimbria andalveus), and (4) parahippocampal gyrus WM and GM.Except for whole brain, structural volumes were deter-mined on two separate occasions. Segmentations ofbrains were completed blindly without investigatorknowledge of the clinical status of the animal (i.e. DAtoxicosis or non-DA etiology).

For the live versus postmortem-fixed comparisons,3D reconstructions and volumes of brain structureswere determined for (1) GM and WM of the total, leftand right cerebral hemispheres, (2) hippocampus and(3) parahippocampal gyrus WM and GM. Structure vol-umes were also determined on two separate occasions.CSF and ventricle volumes were not reported becausefluid in the brain would have been replaced with for-malin. GM and WM of the entire brain and cerebellumfor the sea lions ‘Fairbanks’, ‘Tiki’ and ‘Rupert’ werenot reported because a portion of the cerebellum andbrainstem were removed for chemical analysis.

Data analysis. For live versus 2, 24, and 48 h post-mortem-intact comparisons, the mean volume (cm3 ormm3) and SD of these structures were reported foreach sea lion (i.e. ‘Gratitude’, ‘Kirina’ and ‘Barlich’) ateach imaging time point (i.e. live and 2, 24 and 48 hpostmortem). The percentage of cerebral hemisphereoccupied by the left or right hippocampus was calcu-lated by dividing the structure volume (i.e. from nativeT2-weighted images) by the sum of the left or rightcerebral WM and GM volumes (i.e. from processedPD-weighted images) multiplied by 100. For each sealion at each postmortem-intact time point, the percent-age of the live volume for each structure was calcu-lated by dividing the postmortem volume by the livevolume multiplied by 100. The average percentage (%)and SD of the live volume of the 3 sea lions werereported for each structure. Data analysis for the liveversus postmortem-fixed comparisons followed a simi-lar approach as postmortem-intact analysis. Statisticalanalysis was not completed because of low samplesizes (i.e. n = 3 for live versus postmortem-intact com-parisons; n = 6 for live versus postmortem-fixed com-parisons).

RESULTS AND DISCUSSION

MR imaging: live versus 2, 24 and 48 h postmortem-intact

The FLASH sequences of sea lions ‘Gratitude’,‘Kirina’ and ‘Barlich’ revealed changes in image qual-

246


ity of the brain as postmortem time increased, whichindicated that this acquisition protocol was not usefulas a postmortem diagnostic tool. The images thatwere acquired while the sea lions were alive illus-trated a distinct contrast between GM and WM ofthe brain (Fig. S1A in the supplement available atwww.int-res.com/articles/suppl/d091p243_supp.pdf).As postmortem time increased, the contrast betweenGM and WM was lost. The images acquired alive andpostmortem-intact revealed a high-intensity signal inthe posterior pituitary gland, which was present in allsea lions (Fig. S1). This finding is normal and com-monly seen on T1-weighted images (Colombo et al.1987). Evidence suggests that the high-intensity sig-nal originates from lipids or lipid-like material, whichmay be linked to hormone production (Colombo et al.1987).

Because of the lack of contrast between the GM andWM in the postmortem-intact images, the FLASHsequence was not a reliable MRI sequence for post-mortem evaluation of the brain. Thus, our focus shiftedto using PD- and T2-weighted images acquired duringthe TSE sequences. PD-weighted transverse imagesacquired live and postmortem-intact revealed excel-lent contrast between GM and WM of the cerebralhemispheres and cerebellum (Fig. S2 in the supple-ment). In the PD- and T2-weighted transverse imagesof the live sea lions, the mesencephalic aqueduct wasidentified but characterized by a signal void (Figs. S2 &S3 in the supplement). The lack of CSF signal in theaqueduct represents a flow artifact secondary to pul-satile CSF flow (Feinberg & Mark 1987, Malko et al.1988, Lisanti et al. 2007). However, on postmortembrain images, CSF signal in the mesencephalic aque-duct and vasculature was present in the PD- and T2-weighted transverse images due to lack of flow. Asexpected, the transverse PD-weighted images werebetter for visualizing WM and GM contrast (Fig. S2),while the T2-weighted images were better for identify-ing CSF (Fig. S3).

In the images acquired live and postmortem-intact,the structures of the hippocampal formation were bestvisualized in the oblique PD- and T2-weighted images,as noted in the non-DA neurologically normal Califor-nia sea lion ‘Kirina’ (Figs. 1 & S4 in the supplement).The cornu ammonis or hippocampus proper and thesubiculum were easily visualized up to 48 h post-mortem (Figs. 1 & S4). The parahippocampal gyruswas easily recognized in the postmortem MR images.WM tracts of the hippocampus, the alveus and fimbria,and WM tracts of the parahippocampal gyrus wereidentifiable in the postmortem images; however, theboundaries between WM and GM of these structureswere more visible in the live and 2 h postmortemimages. Thus, imaging brains postmortem, up to 48 h

after death, provides adequate images of the hippo-campus and surrounding structures.

The sea lions ‘Gratitude’ and ‘Barlich’ exhibited clin-ical signs of DA toxicosis, and their feces containeddetectable concentrations of DA (Table 1). The T2-weighted images derived from scanning the brain of‘Barlich’ alive revealed atrophy of the right hippocam-pus (i.e. cornu ammonis) with ex vacuo ventriculo-megaly (i.e. enlargement of the ventral horn of theright lateral ventricle) (Fig. 2A). These images alsorevealed thinning of the right parahippocampal gyrus.These findings were also observed in the T2-weightedimages obtained postmortem (Fig. 2B–D). However,the WM of the subiculum and parahippocampal gyruswas not as well defined and blended in with the GM.CSF surrounding the hippocampus was more preva-lent in the postmortem images.

Brain volumes: live versus 2, 24 and 48 hpostmortem-intact

Volumes of all brain structures derived from imagingsea lions alive are reported in Table 3. Segmentation ofthe transverse T2-weighted images was used to delin-eate the brain surface and calculate whole brain vol-ume for sea lions imaged alive and 2, 24 and 48 h post-mortem-intact (Fig. S5 in the supplement). The wholebrain volumes that were determined from imagesacquired alive were similar to the volumes obtainedfrom images acquired postmortem-intact (Fig. S5). Thevolumes of total CSF and ventricles were determinedfrom segmentations of the transverse T2-weightedimages after digital removal of nearby blubber, mus-cle, skull and other head anatomy (Fig. S5, Table S1in the supplement). Generally, the volumes of CSFincreased at 2 h postmortem, then decreased at 24 and48 h after death (Fig. S5, Table S1 in the supplement).This pattern is best explained by the fact that at 2 hpostmortem the diffusion artifact caused by blood flowand CSF flow in a live animal is not present once deathoccurs; therefore, the volume that characterizes thehigh signal intensity of fluid increases, whereas at 24and 48 h postmortem, the volume of CSF decreasesbecause leakage occurs. The general pattern for ven-tricular volume was a decrease from live to 48 h post-mortem, which was most probably due to CSF leakage(Fig. S5, Table S1).

The volumes of GM and WM of the brain were esti-mated from segmentations of the transverse PD-weighted images (Fig. S5, Table S1). The GM andWM volumes of the entire brain, cerebral hemi-spheres and cerebellum determined from images ac-quired alive were, for the most part, similar to the vol-umes determined from images acquired postmortem

247

Dis Aquat Org 91: 243–256, 2010

(Fig. S5, Table S1). Thus, imaging brains postmortem-intact, up to 48 h after death, does not have any ob-vious effect on GM and WM volumes, but does affectCSF volumes.

Volumes of the left and right hippocampus and asso-ciated structures were determined by manual seg-mentation of oblique T2-weighted images (Fig. S6 inthe supplement). The T2-weighted images were usedbecause they were better at highlighting fluid struc-

tures surrounding the hippocampus compared withthe PD images (Figs. 1 & S4). These fluid structuresserved as boundaries of the hippocampus and weredefined by higher signal intensities.

For ‘Kirina’ (i.e. non-DA, neurologically normalfemale sea lion), volumes of the left and right hip-pocampus were approximately equal (Table 3, Fig. 3).In addition, the left and right parahippocampal gyriwere approximately equal in volume (Tables 3 & S2).

248

Fig. 1. Zalophus californianus. T2-weighted magnetic resonance images acquired in the oblique plane of a California sea lion(‘Kirina’) that was imaged (A) alive, (B) 2 h, (C) 24 h and (D) 48 h postmortem-intact. This sea lion did not show any clinical signsof domoic acid toxicosis. The diffusion artifact caused by blood flow in a live animal was not present postmortem (indicated bya yellow arrow). The T2-weighted images were still useful for examination of the hippocampus up to 48 h postmortem. Al: alveus;Ca: cornu ammonis; Cca: caudal cerebral artery; Cs: collateral sulcus; Hs: hippocampal sulcus; Lv(v): lateral ventricle

(ventral horn); Pg: parahippocampal gyrus; Sb: subiculum; Tl: temporal lobe. Scale bar = 8.0 cm


Both the volumes of the left and right hippocampuswere 0.84% of the volumes of the left and right cere-bral hemispheres (Fig. 3C–D). Postmortem-intactimaging did not have an effect on the volumes of thehippocampus or the percentages; however, the vol-umes of the ventral horn of the lateral ventriclesdecreased as postmortem imaging time increased

(Table S2 in the supplement). We attribute this findingto CSF leakage. Most importantly, volumes and per-centages of the hippocampus relative to the cerebralhemispheres determined from postmortem imageswere similar to values calculated from images acquiredalive. Histological evaluation indicated that both theleft and right hippocampus were within normal limits.

249

Fig. 2. Zalophus californianus. T2-weighted magnetic resonance images acquired in the oblique plane of a California sea lion(‘Barlich’) that contained domoic acid (DA) in feces and exhibited clinical signs of DA toxicosis. The animal was imaged (A) alive,(B) 2 h, (C) 24 h and (D) 48 h postmortem-intact. The T2-weighted images revealed atrophy of the right hippocampus (indicatedby yellow arrow) with ex vacuo ventriculomegaly. This condition presents as enlargement of the temporal horn of the rightlateral ventricle. These images also revealed thinning of the right parahippocampal gyrus (indicated by orange arrow). The T2-weighted images were still useful for examination of the hippocampus up to 48 h postmortem. See Fig. 1 for abbreviations.

Scale bar = 8.0 cm

Dis Aquat Org 91: 243–256, 2010

The sea lion ‘Gratitude’ exhibited bilateralhippocampal atrophy, which was substan-tiated by volumetric measurements. Eventhough this female California sea lion waslarger than ‘Kirina’ (137 cm compared with117 cm) and exhibited larger total GM andWM volumes, the left and right hippocampiwere smaller both in absolute and relative vol-umes (Table 3, Fig. 3). Volumes of the left andright hippocampi calculated from imagesacquired postmortem-intact were similar tothe volumes determined from images acquiredalive (Fig. 3). Histopathology confirmed thatboth the left and right hippocampi exhibitedlesions consistent with DA toxicosis includingatrophy, neuronal loss and gliosis. Thus, vol-umes of the hippocampus determined fromimages acquired postmortem-intact providedan accurate method in detecting bilateral hippo-campal atrophy.

Sea lion ‘Barlich’ exhibited unilateral, righthippocampal atrophy, which was also substan-tiated by volumetric measurements. This adultfemale was 155 cm in length, much larger than‘Kirina’, and exhibited larger total GM andWM volumes compared with the volumes ob-served for ‘Kirina’. Yet, the right hippocampuswas about half the size of that for ‘Kirina’(Table 3, Fig. 3). The left hippocampus wassimilar in size to the left hippocampus of ‘Ki-rina’. Volumes of the left and right hippocampicalculated from images acquired postmortem-intact were similar to the volumes determinedfrom images acquired alive (Fig. 3). Grosspathologic evaluation of ‘Barlich’ confirmedmoderate atrophy of the right hippocampus(Fig. 4A). Neuronal and granular cell loss inthe right hippocampal pyramidal cell layer anddentate gyrus and moderate gliosis were notedon histological evaluation, confirming expo-sure to DA (Fig. 4B). The left hippocampus waswithin normal limits (Fig. 4C). Thus, the vol-ume difference between the left and righthippocampus determined from imaging corre-lated with the difference seen histologically.Hence, volumes of the hippocampus deter-mined from images acquired up to 48 h afterdeath provided an accurate method in detect-ing bilateral hippocampal atrophy.

MR imaging: live versus postmortem-fixed

The PD-weighted transverse images of thepostmortem-fixed brains revealed excellent

250

Sea

lion

Bra

inL

eft

Rig

ht

Lef

tR

igh

tL

eft

Rig

ht

Lef

tR

igh

tC

ere-

Cer

e-L

eft

Rig

ht

Lef

tR

igh

tL

eft

Rig

ht

Lef

tR

igh

tL

eft

Rig

ht

nam

e(c

m3 )

CS

FC

SF

ven

tri-

ven

tri-

cere

-ce

re-

cere

-ce

re-

bel

lum

bel

lum

hip

po

hip

po

late

ral

late

ral

Hs

Hs

Pg

Pg

Pg

Pg

(cm

3 )(c

m3 )

cles

cles

bra

lb

ral

bra

lb

ral

and

an

d

(mm

3 )(m

m3 )

ven

tri-

ven

tri-

(mm

3 )(m

m3 )

GM

GM

WM

WM

(cm

3 )(c

m3 )

GM

GM

WM

WM

bra

in-

bra

in-

cle

VH

cle

VH

(mm

3 )(m

m3 )

(mm

3 )(m

m3 )

(cm

3 )(c

m3 )

(cm

3 )(c

m3 )

stem

stem

(mm

3 )(m

m3 )

WM

GM

(cm

3 )(c

m3 )

Gra

titu

de

Mea

n34

8.70

19.4

213

.22

3.44

2.61

85.7

790

.94

29.6

530

.48

28.8

346

.75

486.

8245

9.35

343.

7524

5.73

297.

6120

0.32

1305

.54

1405

.64

291.

3826

5.75

SD

1.24

0.59

0.17

0.15

0.50

0.43

0.10

0.17

0.72

0.04

12.4

335

.04

12.7

710

.88

19.3

412

.95

16.0

560

.94

11.5

73.

63

Kir

ina

Mea

n30

1.71

19.9

419

.02

3.76

3.83

64.8

264

.06

27.0

027

.43

25.6

534

.94

771.

2977

2.91

514.

2249

2.82

313.

7934

8.37

1112

.42

1168

.00

351.

0632

4.29

SD

1.06

1.03

0.05

0.07

2.84

2.12

1.12

0.89

1.13

1.45

16.3

817

.79

8.75

7.49

5.49

15.5

427

.98

29.0

76.

4012

.29

Tin

tore

tto

Mea

n41

2.19

29.9

326

.85

3.38

3.73

98.1

597

.59

39.5

040

.51

31.1

543

.09

862.

7145

3.81

283.

3044

3.56

259.

1331

4.42

1604

.33

1450

.17

412.

4342

9.64

SD

2.81

2.35

0.09

0.18

2.74

2.91

0.75

1.06

0.28

1.18

3.45

13.8

13.

2811

.74

9.84

5.52

7.60

7.77

10.8

815

.19

Tik

iM

ean

347.

6426

.62

22.5

42.

302.

3082

.92

84.6

631

.14

32.6

825

.43

39.0

458

3.09

737.

8725

3.77

212.

8821

7.03

185.

4112

07.2

013

18.5

235

8.01

330.

06S

D0.

450.

630.

060.

020.

160.

990.

600.

410.

250.

198.

1127

.45

4.66

6.90

15.1

910

.18

6.56

25.8

95.

703.

45

Fai

rban

ks

Mea

n36

2.89

28.3

121

.58

2.41

2.59

85.2

685

.44

33.2

734

.18

28.1

343

.62

736.

6557

7.85

258.

9032

8.35

199.

0921

1.78

1280

.93

1057

.31

348.

6125

8.16

SD

2.93

2.32

0.10

0.13

1.20

1.08

0.63

0.41

0.44

0.54

10.1

814

.50

22.2

749

.03

13.2

99.

1599

.78

58.6

96.

901.

90

Ru

per

tM

ean

406.

1135

.56

33.1

84.

464.

4289

.21

89.1

437

.01

37.9

430

.12

49.4

471

9.47

731.

5940

3.61

432.

2634

9.22

344.

1813

50.9

613

88.6

225

8.35

284.

74S

D0.

220.

080.

060.

060.

050.

660.

781.

340.

770.

8515

.78

4.40

23.6

712

.29

5.31

18.5

126

.10

15.3

26.

0720

.03

Bar

lich

Mea

n35

7.50

21.3

818

.66

3.22

3.41

80.3

778

.62

33.7

933

.40

31.9

644

.11

752.

4145

7.13

310.

6632

2.62

300.

8934

7.76

1352

.97

1049

.15

382.

0628

3.07

SD

1.65

1.35

0.14

0.13

0.22

0.91

1.02

0.35

0.52

0.29

58.6

923

.99

6.39

6.39

31.5

923

.65

34.1

88.

468.

6317

.78

Tab

le 3

.Zal

oph

us

cali

forn

ian

us.

Bra

in v

olu

mes

of

Cal

ifor

nia

sea

lion

s d

eter

min

ed f

rom

imag

ing

ali

ve. C

SF

: cer

ebro

spin

al f

luid

; WM

: wh

ite

mat

ter;

GM

: gra

y m

atte

r; h

ipp

o:h

ipp

ocam

pu

s (t

he

volu

me

of t

he

hip

poc

amp

us

incl

ud

ed t

he

corn

u a

mm

onis

, d

enta

te g

yru

s, s

ub

icu

lum

, fi

mb

ria

and

alv

eus)

; V

H:

ven

tral

hor

n;

Hs:

hip

poc

amp

al s

ulc

us

(th

evo

lum

e of

th

e h

ipp

ocam

pu

al s

ulc

us

incl

ud

ed C

SF

of

sulc

us

and

CS

F o

f tr

ansv

erse

fis

sure

of

bic

hat

); P

g:

par

ahip

poc

amp

al g

yru

s (t

he

volu

me

of P

g G

M m

ay h

ave

con

tain

ed G

M o

f th

e su

bic

ulu

m)


contrast between GM and WM, and were similar to thecontrast seen in live images (Fig. S7 in the supple-ment). However, in some cases, the images of the for-malin-fixed brains were flattened in the dorsal–ventraldirection, most probably related to the weight of thebrain (Fig. S7). As expected, the PD-weighted imagesof the postmortem-fixed brains were better for visual-izing WM and GM contrast than the T2-weightedimages. The elapsed time from death to fixation (up to51 h) did not affect the WM and GM contrast.

Similar to MR images acquired live, the oblique T2-weighted images of postmortem-fixed brains were best

for visualizing the hippocampus (Fig. S8 in the supple-ment). However, the boundaries of the hippocampuswere less distinct in these images compared with theboundaries in the images acquired alive (Fig. S8). Thecornu ammonis or hippocampus proper and the subi-culum were recognized; however, in some cases, theboundaries were blurred (Fig. S8D), while in othercases the boundaries remained distinct (Fig. S8F).These patterns were observed for other structures,including the parahippocampal gyrus, as well as WMtracts of the hippocampus, the alveus and fimbria andWM tracts of the parahippocampal gyrus. These ef-

251

0 200 400 600 800 1000

Kirina

Gratitude

Barlich

Volume (mm3) of left hippocampus Volume (mm3) of right hippocampus

0.00 0.20 0.40 0.60 0.80 1.00

Kirina

Gratitude

Barlich

% left hippocampus of left cerebralhemisphere

% right hippocampus of right cerebralhemisphere

0 200 400 600 800 1000

Kirina

Gratitude

Barlich

0.00 0.20 0.40 0.60 0.80 1.00

Kirina

Gratitude

Barlich

0 20

Live

2 h postmortem

24 h postmortem

48 h postmortem

40 60 80 100 120

Lefthippocampus

Righthippocampus

Average % of live volume

C) D)

E

C D

A B

Fig. 3. Zalophus californianus. Volumes (mm3) of the left andright hippocampus from images acquired alive and post-mortem-intact of sea lions ‘Kirina’, ‘Barlich’ and ‘Gratitude’.Volumes of (A) left and (B) right hippocampus. Percent of thecerebral hemisphere occupied by the (C) left and (D) righthippocampus. The percentage of cerebral hemisphere occu-pied by the left or right hippocampal structure was calculatedby dividing the structure volume (i.e. from native T2-weighted images) by the sum of the left or right cerebral WMand GM volumes (i.e. from processed PD-weighted images)multiplied by 100. (E) Average percent (%) of live volume forthe left and right hippocampi. This value was calculated bydividing the postmortem volume by the live volume andmultiplying by 100% for ‘Kirina’, ‘Barlich’ and ‘Gratitude’.The mean and SD of these 3 sea lions were reported. Struc-tural volumes were determined twice, which was the source

of the error bars

Dis Aquat Org 91: 243–256, 2010

fects were not associated with the elapsed time fromdeath to fixation. Thus, imaging formalin-fixed brainsprovides some information on brain structure; how-ever, the images of the hippocampus and surroundingstructures were of poorer quality (i.e. with regards tocontrast of WM and GM) than images acquired aliveand postmortem-intact.

The sea lions ‘Fairbanks’, ‘Tiki’, ‘Tintoretto’, ‘Barlich’and ‘Gratitude’ exhibited clinical signs of DA intoxica-tion, but only ‘Gratitude’ and ‘Barlich’ containeddetectable concentrations of DA in feces (Table 1).Scanning the postmortem-fixed brain, in some cases,was still useful in diagnosing hippocampal atrophy(Fig. S8). For example, the T2-weighted images de-rived from scanning the formalin-fixed brain of ‘Bar-lich’ revealed atrophy of the right hippocampus (i.e.

cornu ammonis) (Fig. S8F). However, in other cases,scanning the formalin-fixed brain provided less thanoptimal images of the hippocampus, and it was difficultto detect hippocampal changes (Fig. S8B,D).

Brain volumes: live versus postmortem-fixed

Volumes of all brain structures derived from imagingsea lions alive are reported in Table 3. The volumes ofcerebral GM and WM of the postmortem-fixed brainswere estimated from segmentations of the transversePD-weighted images (Fig. 5, Table S3 in the supple-ment). GM and WM volumes of the cerebellum for sealions ‘Rupert’, ‘Tiki’ and ‘Fairbanks’ were not deter-mined because a portion of the cerebellum had been

252

Fig. 4. Zalophus californianus. Gross and histological images of ‘Barlich’, a California sea lion that contained domoic acid (DA) infeces and exhibited clinical signs of DA toxicosis. (A) The gross image revealed atrophy of the right hippocampus with ex vacuoventriculomegaly. The left hippocampus was normal. Dotted squares encompass the left and right hippocampi and para-hippocampal gyrus. Scale bar = 1cm. (B) H&E stain showing neuronal and granular cell loss in the right hippocampal pyramidalcell layer and dentate gyrus and moderate gliosis. (C) H&E stain of the normal left hippocampus, which is much larger than the

right hippocampus. CA: cornu ammonis; Dg: dentate gyrus. Scale bars (in B,C) = 1 mm


removed for chemical analysis. The GM and WM vol-umes of the cerebral hemispheres determined fromimages of formalin-fixed brains were, for the most part,similar to the volumes obtained from images acquiredalive.

Volumes of the left and right hippocampi and asso-ciated structures of the postmortem-fixed brains weredetermined by manual segmentation of oblique T2-weighted images. Imaging fixed brains did not havean obvious effect on the volumes of the hippocampior the percentages compared with live imaging(Fig. 5). Volumetric calculations of the hippocampifrom postmortem-fixed images provided an accuratemethod in detecting unilateral and bilateral hip-pocampal atrophy. The sea lion ‘Gratitude’ exhibiteda decrease in volume of the left and right hippocam-pus compared with the volumes measured in ‘Rupert’(a sea lion that did not show any clinical signs of DAtoxicosis). Histopathology confirmed the findings for‘Gratitude’, as previously discussed. On histologicalevaluation of ‘Rupert’s’ brain, the hippocampi werenormal. The brains of ‘Barlich’, ‘Tintoretto’ and ‘Fair-banks’ showed a decrease in volume of the right hip-pocampus, while that of ‘Tiki’ revealed a loss in vol-ume of the left hippocampus (Fig. 5). Histopathologyconfirmed the MRI findings. Lesions consistent withDA poisoning, including hippocampal pyramidal andgranular cell loss and gliosis, were only present in theright hippocampus of ‘Barlich’, ‘Tintoretto’ and ‘Fair-banks’ and in the left hippocampus of ‘Tiki’. Thus,the volume differences between the left and righthippocampus determined from imaging formalin-fixed brains correlated with the differences observedhistologically.

Understanding hippocampal atrophy and domoicacid toxicosis

In the present study, volumes of the hippocampi fromimages acquired live, postmortem-intact, and post-mortem-fixed provided an accurate method in detect-ing unilateral and bilateral hippocampal atrophy asso-ciated with DA toxicosis. However, it was important toexpress the volume of the left or right hippocampusrelative to the volumes of the left or right cerebralhemispheres (i.e. percent hippocampus) because thehippocampus grows in size from pups to adults. For thefemale yearling (‘Kirina’) that did not exhibit any clini-cal signs of DA toxicosis, the volumes of the left andright hippocampi were 0.84% of the volumes of the leftand right cerebral hemispheres. Whether this percent-age can serve as a baseline normal value remains inquestion and will require the analysis of more normalcases. Thus, future research should focus on how the

volume of the normal hippocampus varies with sex,length and age of the sea lion.

In addition, it is worthwhile to mention that usingpercentage hippocampus (relative to the left or rightcerebral hemisphere, respectively) as a means toassess the degree of hippocampal atrophy associatedwith DA toxicosis does need further exploring be-cause atrophy of other brain structures within thecerebral hemispheres may be associated with DA tox-icosis. Silvagni et al. (2005) showed that sea lions withDA toxicosis exhibited necrotic neurons throughoutmuch of the cortical and thalamic limbic system, in-cluding olfactory bulb, pyriform lobe and rostral thal-amic nuclei. Voxel-based morphometry studies inhumans with temporal lobe epilepsy (TLE) haveshown a reduction in volume of many brain regionsrelative to those of healthy controls (Keller & Roberts2008). These structures included the cerebellum, mid-brain, claustrom, globus pallidus, putamen, caudatenucleus, thalamus, occipital lobe, parietal lobe, insula,cingulate gyrus, orbital frontal lobe, dorsal frontallobe, fornix, Heschl’s gyrus, superior temporal gyrus,middle temporal gyrus, inferior temporal gyrus, fusi-form gyrus, parahippocampal gyrus, entorhinal cor-tex, amygdala and hippocampus (Keller & Roberts2008). Epilepsy in California sea lions has beenrecently associated with chronic consequences of sub-lethal exposure to DA (Goldstein et al. 2008). Thus,future research should investigate how DA affectsother brain structures besides the hippocampus.These investigations would be helpful in determiningif percentage hippocampus is the ‘gold standard’ inassessing the chronic consequences of DA toxicosis inCalifornia sea lions.

It is curious that some sea lions exhibited unilateralatrophy of the hippocampus, whereas other individu-als showed bilateral atrophy. Goldstein et al. (2008)also observed the occurrence of unilateral brain lesionsin California sea lions exposed to sublethal levels ofDA. In that study, more sea lions exhibited unilateraldamage than bilateral damage, and right-sided atro-phy was most common (Goldstein et al. 2008). Themechanism involved in the pathology of unilateral ver-sus bilateral hippocampal atrophy is unclear but maybe linked to seizure origin. In sea lions, it is possiblethat atrophy is ipsilateral to the seizure focus. Thus, ifseizures occur in both hemispheres, then the animalwould exhibit bilateral atrophy; if seizures originate inthe right hemisphere, then the sea lion would exhibitright hippocampal atrophy. In humans with temporallobe epilepsy, volume reduction of the hippocampusipsilateral to the side of the seizure is the primary find-ing in voxel-based morphometry studies (as reviewedin Keller & Roberts 2008). This finding supports ourproposed mechanism.

253

Dis Aquat Org 91: 243–256, 2010

CONCLUSIONS

This investigation presents a comparison of MRIquality and volumes of brain structures obtained andderived from imaging the brain of California sea lionslive, postmortem-intact, and postmortem-fixed. PD-

and T2-weighted images provided sufficient quality(i.e. signal to noise and contrast of WM and GM) toview the brain up to 48 h postmortem, which allowedthe determination of volumes of brain structures thatwere comparable with live imaging. Volumes of thehippocampus from postmortem-intact images (up to

254

0 50 100 150 200 250

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

Cerebral gray matter volume (cm3) Cerebral white matter volume (cm3)0 20 40 60 80 100

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

0 200 400 600 800 1000

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

Volume of left hippocampus (mm3) Volume of right hippocampus (mm3)

0 200 400 600 800 1000

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

0.0 0.2 0.4 0.6 0.8 1.0

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

% left hippocampus of left cerebralhemisphere

0.0 0.2 0.4 0.6 0.8 1.0

Fairbanks

Tiki

Rupert

Tintoretto

Barlich

Gratitude

% right hippocampus of right cerebralhemisphere

A B

C D

E F

Live

Formalin

Fig. 5. Zalophus californianus. Volumes (mm3 or cm3) of brain structures from live and formalin-fixed brains of 6 California sea lions. (A) Cerebral gray matter volume. (B) Cerebral white matter volume. (C,D) Volumes of (C) left and (D) right hippo-campus. (E,F) Percent of the cerebral hemisphere occupied by (E) the left and (F) the right hippocampus. Fixation of the brainoccurred at 2 h after death for ‘Fairbanks’, 3 h after death for ‘Tiki’, 4 h after death for ‘Rupert’, 23 h after death for ‘Tintoretto’,48 h after death for ‘Barlich’ and 51 h after death for ‘Gratitude’. Structural volumes were determined twice, which was the source

of the error bars


48 h after death) provided an accurate method fordetecting unilateral and bilateral atrophy that wascomparable with live imaging. Imaging postmortem-fixed brains provided some information on brain struc-ture; however, the images of the hippocampus and sur-rounding structures were of poorer quality thanimages acquired from live animals and postmortem-intact. Despite these problems, volumes of cerebralGM and WM, as well as hippocampus, were similar tovolumes calculated from images acquired alive andwere sufficient to detect hippocampal atrophy.

Thus, postmortem MRI scanning (either intact or for-malin-fixed) with volumetric analysis can be used toinvestigate the acute, chronic and possible develop-mental effects of DA on the brain of California sealions. While MRI will never replace histological evalu-ation, it does permit broad coverage of the brain withvery thin sections, which can be helpful in localizinglesions. In addition, volumetric calculations of brainstructures add a powerful quantitative dimension.These advantages can increase the chances of a defin-itive quantitative diagnosis. This approach can also beused to investigate the effects of other biological,chemical and physical agents on wild California sealions, other marine mammals and wildlife in general.

Acknowledgements. The authors thank all the staff and vol-unteers at The Marine Mammal Center. They also thank Drs.H. Harris and F. Nutter for their veterinary assistance and Dr.G. Schneider for his expertise in neurobiology. The authorsalso thank J. Bloom and Dr. C. Kruse-Elliot for MRI and theIAMs imaging center, Redwood City, California for theirassistance. This study was supported under a Subaward withthe University Corporation of Atmospheric Research (UCAR)under Grant No. NA06OAR4310119 (Training Tomorrow’sEcosystem and Public Health Leaders Using Marine Mam-mals as Sentinels of Oceanic Change) with the NationalOceanic and Atmospheric Administration (NOAA), USDepartment of Commerce. We also acknowledge the partnersof this training grant: University of California Davis WildlifeHealth Center, The Marine Mammal Center and NorthwestFisheries Science Center (NWFSC). Supplemental fundingwas provided by The Marine Mammal Center, the USNational Marine Fisheries Service and Dr. D. Mann, Collegeof Marine Science, University of South Florida.

LITERATURE CITED

Colombo N, Berry I, Kucharczyk J, Kucharczyk W and others(1987) Posterior pituitary glands appearance on MRimages in normal and pathologic state. Radiology 165:481–485

Conrad PA, Miller MA, Kreuder C, James ER and others(2005) Transmission of toxoplasma: clues from the study ofsea otters as sentinels of Toxoplasma gondii flow into themarine environment. Int J Parasitol 35:1155–1168

Costa LG, Giordano G (2007) Developmental neurotoxicity ofpolybrominated diphenyl ether (PBDE) flame retardants.Neurotoxicology 28:1047–1067

Dua-Sharma S, Sharma KN, Jacobs HL (1970) The caninebrain in stereotaxic coordinates. MIT Press, Cambridge,MA

Feinberg DA, Mark AS (1987) Human brain motion and cere-brospinal fluid circulation demonstrated with MR velocityimaging. Radiology 163:793–799

Goldstein T, Mazet JAK, Zabka TS, Langlois G and others(2008) Novel symptomatology and changing epidemiol-ogy of domoic acid toxicosis in California sea lions (Zalo-phus californianus): an increasing risk to marine mammalhealth. Proc Biol Sci 275:267–276

Greig DJ, Gulland FMD, Kreuder C (2005) A decade of liveCalifornia sea lion (Zalophus californianus) strandingsalong the Central California coast: causes and trends,1991–2000. Aquat Mamm 31:11–22

ICVGAN (International Committee on Veterinary GrossAnatomical Nomenclature) (2005) Nomina anatomicaveterinaria, 5th edn. World Association of VeterinaryAnatomists, Ithaca, NY

Jepson PD, Arbelo M, Deaville R, Patterson IAP and others(2003) Gas-bubble lesions in stranded cetaceans. Nature425:575–576

Keller SS, Roberts N (2008) Voxel-based morphometry oftemporal lobe epilepsy: an introduction and review of theliterature. Epilepsia 49:741–757

Lisanti C, Carlin C, Banks KP, Wang D (2007) Normal MRIappearance and motion-related phenomena of CSF.Am J Roentgenol 188:716–725

Malko JA, Hoffman JC, McClees EC, Davis PC, Braun IF(1988) A phantom study of intracranial CSF signal loss dueto pulsatile motion. Am J Neuroradiol 9:83–89

Marino L, Murphy TL, Deweerd AL, Morris JA and others(2001a) Anatomy and three-dimensional reconstructionsof the brain of the white whale (Delphinapterus leucas)from magnetic resonance images. Anat Rec 262:429–439

Marino L, Murphy TL, Gozal L, Johnson JI (2001b) Magneticresonance imaging and three dimensional reconstructionsof the brain of a fetal common dolphin, Delphinus delphis.Anat Embryol 203:393–402

Marino L, Sudheimer KD, Murphy TL, Davis KK and others(2001c) Anatomy and three-dimensional reconstructionsof the brain of a bottlenose dolphin (Tursiops truncatus)from magnetic resonance images. Anat Rec 264:397–414

Marino L, Sudheimer K, Sarko D, Sirpenski G, Johnson JI(2003a) Neuroanatomy of the harbor porpoise (Phocoenaphocoena) from magnetic resonance images. J Morphol257:308–347

Marino L, Sudheimer K, Pabst DA, McLellan WA, Johnson JI(2003b) Magnetic resonance images of the brain of adwarf sperm whale (Kogia simus). J Anat 203:57–76

Marino L, Sudheimer K, McLellan WA, Johnson JI (2004a)Neuroanatomical structures of the spinner dolphin (Ste-nella longirostris orientalis) brain from magnetic reso-nance images. Anat Rec 279A:601–610

Marino L, Sherwood CC, Delman BN, Tang CY, Naidich TP,Hof PR (2004b) Neuroanatomy of the killer whale (Orcinusorca) from magnetic resonance images. Anat Rec 281A:1256–1263

Montie EW, Schneider GE, Ketten DR, Marino L, Touhey KE,Hahn ME (2007) Neuroanatomy of the subadult and fetalbrain of the Atlantic white-sided dolphin (Lagenorhyn-chus acutus) from in situ magnetic resonance images.Anat Rec 290:1459–1479

Montie EW, Schneider GE, Ketten DR, Marino L, Touhey KE,Hahn ME (2008) Volumetric neuroimaging of the brain ofthe Atlantic white-sided dolphin (Lagenorhynchus acu-tus). Anat Rec 291:263–282

255

Dis Aquat Org 91: 243–256, 2010

Montie EW, Pussini N, Schneider GE, Battey TWK, DennisonS, Barakos J, Gulland F (2009) Neuroanatomy and vol-umes of brain structures of a live California sea lion (Zalo-phus californianus) from magnetic resonance images.Anat Rec 292:1523–1547

Nolte J, Angevine JB (2000) The human brain in photographsand diagrams, 2nd edn. Mosby, St. Louis, MO

Ramsdell JS, Zabka TS (2008) In utero domoic acid toxicity: afetal basis to adult disease in the California sea lion (Zalo-phus californianus). Mar Drugs 6:262–290

Scholin CA, Gulland F, Doucette GJ, Benson S and others(2000) Mortality of sea lions along the central Californiacoast linked to a toxic diatom bloom. Nature 403:80–84

Sharlin DS, Bansal R, Zoeller RT (2006) Polychlorinated

biphenyls exert selective effects on cellular composition ofwhite matter in a manner inconsistent with thyroid hor-mone insufficiency. Endocrinology 147:846–858

Silvagni PA, Lowenstine LJ, Spraker T, Lipscomb TP, GullandFMD (2005) Pathology of domoic acid toxicity in Californiasea lions (Zalophus californianus). Vet Pathol 42:184–191

Tiedeken JA, Ramsdell JS (2010) Zebrafish seizure modelidentifies p,p’- DDE as the dominant contaminant of fetalCalifornia sea lions that accounts for synergistic activitywith domoic acid. Environ Health Perspect 118:545–551

Zoeller RT, Dowling ALS, Herzig CTA, Iannacone EA,Gauger KJ, Bansal R (2002) Thyroid hormone, braindevelopment, and the environment. Environ Health Per-spect 110:355–361

256

Editorial responsibility: Michael Moore,Woods Hole, Massachusetts, USA

Submitted: April 26, 2010; Accepted: June 29, 2010Proofs received from author(s): August 30, 2010

disease of aquatic organisms 91:243

Documents