AN MRI STUDY OF SEPTI PELLUCIDI IN RELATION TO HIPPOCAMPUS AND FORNIX IN SCHIZOPHRENIA

Davidson, C. A.(1), Levitt, J. J.(1,2), Kuroki, N.(2), Niznikiewicz, M.(2), McCarley, R. W.(2), and Shenton, M. E.(1,2)

(1) Psychiatry Neuroimaging Laboratory, Brigham and Women’s Hospital, Harvard Medical School(2) Clinical Neuroscience Division, Laboratory of Neuroscience, Boston VA HealthCare System-Brockton Division, Department of Psychiatry, Harvard Medical School

Background

Methods

Results

The septum pellucidum (SP) is a midline limbic structure that divides the lateral ventricles between the corpus callosum and fornix. Cavum septi pellucidi (CSP) is a relatively common neurodevelopmental variant in which the laminae of the SP do not fuse completely. Large CSP, small hippocampus, and fornix abnormalities have all been reported in schizophrenia (SZ) and related disorders, and are likely associated with alterations in neurodevelopment.

MR images from 24 male, chronic SZs and 30 normal controls (NC), group matched for parental socio-economic status, handedness, age and intra-cranial contents (ICC) were acquired using an SPGR sequence on a 1.5-T GE scanner. CSP length was measured using Kwon et al.’s (1998) criteria and CSP width and SP length were measured similarly to Born et al. (2004). Fornix fractional anisotropy (FA), mean diffusivity (Dm), and cross-sectional area as well as hippocampus volume (HC) measurements were made as previously reported by Kuroki et al. (2006).

References

Conclusions & Further Directions

Top Left, key middle right: Coronal SP (Andy & Stephan, 1968)

Middle Left: MR images of medium-sized CSP Bottom Left: MR images of normal lack of CSP (Born et al, 2004)

Bottom Right:: SP Anomaly Cartoon (Born et al, 2004)

Fornix ROI on DTI map (Kuroki et al, 2006) Fractional Anisotropy model & 3D DTI image (Kuroki et al, 2006)

CSP width and length in horizontal section (Filipovic,Teofilovski-Parapid,2004) CSP ROI, linear CSP width and length. (ROI is

actually drawn on coronal plane)Horizontal SP measure: Yellow, parallel to AC-PC line: white, blue dots for AC and PC.

Diagonal SP measure: Light Blue.

Coronal & Sagittal 3D renderings; coronal HC ROI (Dickey, et al, 2007)

CSP measures did not differ between diagnostic groups, although due to the small sample size and rarity of CSP abnormalities a Type II error cannot be precluded. SP length relative to ICC was significantly larger in SZ than NC (p=0.031). In SZ, fornix FA showed trend-level positive correlation with CSP size and a significant negative correlation with SP length (p<0.01). Also in SZ, right hippocampus volume correlated positively with CSP size (p<0.05), and bilateral hippocampus volume correlated negatively with SP length (p<0.01). There were no such correlations in controls.

Pearson’s Correlation: -0.673, p=0.000 Spearman’s Rho: -0.586, p=0.003

SP length measures, absolute and relative to ICC, showed a strong negative correlation with fornix FA in the SZ group but not in the NC group.

50

Pearson’s: 0.469, p=0.021 Spearman’s Rho: 0.413, p=0.045

In both groups, fornix FA correlated positively with the angle between the two SP length ROIs. This angle represents a relationship between the genu of the corpus callosum and the most superior aspect of the fornix.

Pearson’s: -0.432, p=0.035 Spearman’s Rho: -0.587, p=0.003

In the SZ group, SP length measures, absolute and relative to ICC, showed a strong negative correlation with bilateral HC measurements, absolute and relative. No such correlations were found in the NC group.

Of SP and CSP measurements, only SP length showed a group difference.

Contrary to a large body of schizophrenia research, this sample showed no differences by diagnosis in any CSP measure. This is not very surprising considering the normal prevalence of SP anomalies, and the large CSPs found in the NC group are only slightly more than the expected 10%.

Excluding these NC outliers did not change any reported results.

Pearson’s: 0.383, p=0.065

CSP length relative to SP length and to ICC showed a trend towards positive correlation with fornix FA in the SZ group but not NC. Spearman’s correlations only showed trends when cases with 0 CSP were excluded.

Spearman’s Rho: 0.473, p=0.019 Pearson’s: 0.367, p=0.078

CSP size relative to ICC and to SP length correlated positively with right HC volume relative to ICC. This correlation was only trend level using absolute CSP measures.

Increased prevalence of CSP has been robustly reported in SZ-spectrum disorders. However, functional and neurodevelopmental implications (primary/secondary) and morphology of SP abnormalities aren’t well understood.

This is in part due to the small size of the SP, which is subject to significant partial volume confounds even with the strongest magnets and best sequences. Also, the human SP remains something of an enigma in basic neuroscience. However, divergent and vaguely-delineated methodologies plague the field.

Recent research has begun to explore the detailed morphology and linear parameters of CSP. Nopolous and colleagues’ (1998) oft-cited method involved classifying CSP as having “normal” or “abnormal” length. Born and colleagues (2004) measured CSP as a percentage of SP length. Filipovic and Teofilovski-Parapid (2004) suggest both width and length as predictive aspects of symptomatic cava, while Galarza and colleagues (2004) suggest three categories corresponding to embryonic stages, delineating “severity” rather than “abnormality.”

Each of these methods for measurement and interpretation are valid for different reasons, but with some exceptions, little has been done to reconcile disparate reports or to go beyond passive observation.

Correlations between SP abnormalities and the temporal lobe have been examined, (Dickey et al, 2007; Kwon et al, 1998; Galarza et al, 2004), but relationships to midline structures haven’t been exploredthoroughly.

Andy, OJ, Heinz, S (1968): The septum in the human brain. J Comp Neur 133:383-410.Born, CM, Meisenzahl, EM, Frodl, T, Pfluger, T, Reiser, M, Moller, HJ, Leinsinger, GL (2004): The septum pellucidum and its variants: An MRI study. Eur Arch Psychiatry Clin Neurosci 254:295-302.Dickey, CC, McCarley, RW, Xu, ML, Seidman, LJ, Voglmaier, MM, Niznikiewicz, MA, Connor, E, Shenton, ME (2007): MRI abnormalities of the hippocampus and cavum septi pellucidi in females with schizotypal personality disorder. Schizophrenia Res 89:49-58.Filipovic, B, Teofilovski-Parapid, G (2004): Linear parameters of normal and abnormal cava septi pellucidi: A post-mortem study. Clin Anat17:626-630.Galarza, M, Merlo, AB, Ingratta, A, Albanese, EF, Albanese, AM (2004): Cavum septum pellucidum and its increased prevalence in schizophrenia: A neuroembryological classification. J Neuropsychiatry Clin Neurosci 16:41-46.Kasai, K, McCarley, RW, Salisbury, DF, Onitsuka, T, Demeo, S, Yurgelun-Todd, D, Kikinis, R, Jolesz, FA, Shenton, ME (2004): Cavum septi pellucidi in first-episode schizophrenia and first-episode affective psychosis: an MRI study. Schizophrenia Res 71:65-76. Kuroki, N, Kubicki, M, Nestor, PG, Salisbury, DF, Park, HJ, Levitt, JJ, Woolston, S, Frumin, M, Niznikiewicz, M, Westin, CF, Maier, SE, McCarley, RW, Shenton, ME (2006): Fornix integrity and hippocampal volume in male schizophrenic patients. Biol Psychiatry 60:22-31.Kwon, JS, Shenton, ME, Hirayasu, Y, Salisbury, DF, Fischer, IA, Dickey, CC, Yurgelun-Todd, D, Tohen, M, Kikinis, R, Jolesz, FA, McCarley, RW (1998): MRI study of cavum septi pellucidi in schizophrenia, affective disorder, and schizotypal personality disorder. Am J Psychiatry 155:509-515.Nopoulos, P, Giedd, JN, Andreasen, NC, Rapoport, JL (1998): Frequency and severity of enlarged cavum septi pellucidi in childhood-onset schizophrenia. Am J Psychiatry 155:1074-1079.Rakic, P, Yakovlev, PI (1968): Development of the corpus callosum and cavum septi in man. J Comp Neur 132:45-72.Sheehan, TP, Chambers, RA, Russell, DS (2004): Regulation of affect by the lateral septum: implications for neuropsychiatry. Brain Res Rev 46:72-117

.

Top right:: brain model, SP highlighted (www..sci.uidaho.edu/med532/septum.htm)

HC ROI dorsal-ventral coronal 3D rendering, sagittal rendering, and coronal manual tracing

(Dickey, et al, 2007)

The present study is limited by a SZ sample with relatively small CSP. We found new relationships between the septum pellucidum and schizophrenia. Post-mortem, MRI, and DTI methods should be utilized on larger samples to begin understanding the relevance of midline structures to schizophrenia spectrum disorders and related psychopathology.