Human Studies

Automated brain tumor and vessel segmentation would enable multiple clinical applications ranging from better guidance of endovascular procedures to new methods of assessing tumor treatment response. No effective, automated approach now exists for segmentation of either set of structures.

Fig. 18 Normalization of vessel shape. Top row: T1 GAD slices at times 1 and 2 and segmented vessels with tumor at time 1. Second row: Vessels at time 1. Third row: Vessels at time 2.

Aim 1: To refine and automate our current vessel and brain tumor segmentation methods from high quality MR and MRA images of brain.

Aim 2: To develop and evaluate a new method for fully deformable, diffeomorphic, multi-modal atlas formation that simultaneously considers tissue structure and vasculature for the localization of common vascular regions across subjects.

Aim 3: To collect an image database of healthy intracranial vascular anatomy subdivided by age and gender.

Aim 4: To develop effective mathematical and statistical methods of vessel/shape analysis and to determine the efficacy of the approach in assessing both tumor malignancy and tumor response to treatment via clinical studies.

Preliminary results

Fully automated vessel segmentation is difficult under conditions of noise and the presence of non-vessel objects. We have already developed a partially automated method of vessel segmentation that, proceeding from a user-supplied seed point, extracts an image intensity ridge that provides the vessel skeleton followed by an adaptive scale assessment of radius at each skeleton point[57]. The approach allows segmentation of vessels even one voxel wide and is resistant to image noise [57]. Post-processing of the segmented vessels allows creation of detailed, connected vessel trees suitable for many applications[58]. With the ability to segment intracranial vessels, it is now possible to determine how intracranial vessels may be altered before and after treatment. Fig. 18 illustrates the results of vessel normalization following successful treatment. Notice that the vessels become more smooth (arrows) after a successful treatment than that prior to treatment.

Representative Publications:

Aylward S, Jomier J, Vivert C, Ledigarcher V, Bullitt E (2005) Spatial graphs for intracranial vascular network characterization, generation, and discrimination. MICCAI 2005; LNCS 3749: 59-66.

Bullitt E, Gerig G, Pizer S, Aylward SR (2003) Measuring tortuosity of the intracerebral vasculature from MRA images. IEEE-TMI 22:1163-1171.

Bullitt E, Zeng D, Gerig G, Aylward S, Joshi S, Smith JK, Lin W, Ewend MG (2005) Vessel tortuosity and brain tumor malignancy: A blinded study. Academic Radiology 12:1232-1240.

top of page

Principal Investigator: Gregory Essick, Ph.D.

CNS Processes Underlying Pain Regulation and Persistence: Functional Imaging

of CNS Regulatory Processes

Temporomandibular disorders (TMD) are painful musculoskeletal disorders involving the muscles of mastication and/or jaw joint and are three times more common in women of reproductive age than in men of comparable age. One putative explanation for pain experienced in TMD is a dysfunction of mechanisms in the central nervous system (CNS) that normally regulate the intensive, temporal, and spatial dimensions of tactile sensory experience. Psychophysical studies have demonstrated that both experimental and clinical pain reduce tactile perception. Understanding the CNS mechanisms through which pain and innocuous tactile stimuli modulate one another may provide insight into the complexity of chronic pain conditions such as TMD.

The objective of this project is to probe the operation of sensory regulatory mechanisms at the level of the cortex in healthy individuals and then apply that knowledge to investigation of patients with persistent musculoskeletal pain. Our working hypotheses are that patients diagnosed with TMD will exhibit abnormal cortical responses to innocuous tactile and noxious thermal stimuli when presented separately and concurrently. Using fMRI, our current specific aims are the following:

1. In normal (i.e., pain-free) subjects, to determine the location and extent to which concurrent noxious thermal stimulation alters the normal fMRI response to tactile stimulation of the skin.

2. In normal subjects, to determine the location and extent to which concurrent tactile stimulation alters the normal fMRI response to noxious thermal stimulation of the skin.

3. To determine in TMD patients whether the patterns and characteristics of the fMRI responses reflecting central tactile-pain interactions differ systematically from those observed in normal subjects.

Preliminary Results

Preliminary data from normal subjects demonstrate spatial topography in the hemodynamic responses to flutter alone (weak, 26Hz skin vibration), noxious heat alone (49°C, producing a moderate level of pain), and concurrent flutter and heat stimulation. Fig. 1 is grand average data for 10 normal subjects. Consistent with our pilot experiments and previous neuroimaging studies, we see a contralateral SI response and a bilateral SII response to the flutter condition (blue and yellow areas in Fig. 1). Yellow denotes areas in which the addition of concurrent heat did not significantly modify the cortical response to flutter. The addition of a concurrent painful heat stimulus all but wipes out the response to flutter in SI. On the other hand, the addition of heat increased the extent of response to flutter in contralateral SII (green and yellow areas), indicating that tactile responsiveness in SII survives apparent SI inactivation and may be enhanced.

Hollins, M., A. Sigurdsson, and K.A. Morris. Local vibrotactile and pain sensitivities are negatively related in temporomandibular disorders. J Pain 2001; 2(1):46-56.

Roy, E.A., M. Hollins and W. Maixner. Reduction of TMD pain by high-frequency vibration: a spatial and temporal analysis. Pain 2003; 101(3):267-74.

Essick, G.K. Psychophysical assessment of patients with posttraumatic neuropathic trigeminal pain. J Orofacial Pain 2004; 18(4): 345-354.

top of page

Principal Investigator: Gregory Essick, Ph.D. & Grace Baranek, Ph.D.

Unisensory and multisensory interactions in autism

Autism is a pervasive, neurodevelopmental disorder affecting multiple domains of cognitive and sociobehavioral function. Although not currently considered a core feature of autism, abnormal reactions to sensory stimuli are prevalent in autism and are a primary target of clinical intervention strategies despite mixed empirical evidence of their efficacy. It is crucial that any abnormalities in the way autistic individuals process sensory information be elucidated in order to ascertain exactly what aspects of the intervention might be responsible for producing the observed positive outcomes. The intention of this project is to evaluate objective measures of sensory processing abnormalities in autism, across tactile submodalities and across sensory systems. Using psychophysical testing and functional Magnetic Resonance imaging, our current specific aims are the following:

To assess tactile perception in groups of adults with and without autism.
To explore central neural correlates of group differences in pleasant, neutral, and unpleasant (painful) tactile perception.
To investigate the interaction between nonpainful tactile and painful thermal stimulation in groups of adults with and without autism.
To investigate interactions between somatosensory and visual stimulation in groups of adults with and without autism.

Preliminary Results

To date, seven adults with autism and three age- and IQ-matched control participants have completed fMRI protocols aimed at characterizing the neural correlates of single-modality somatosensory perception in autism. While in the scanner, participants experience four types of stimuli: textured surfaces, low frequency vibration, innocuous heat (i.e., warmth), and noxious (painful) heat. The textures are stroked across the forearm; the vibrotactile stimuli are applied to the forearm and the palm one-at-a-time, and all heat stimuli are presented to the palm. Analysis of this data is currently in progress, and preliminary results from our psychophysical studies using similar, unisensory stimuli suggest discrete domains of enhanced tactile perception in autism; participants with autism exhibit superior detection of low frequency vibrations on the forearm but not the palm while demonstrating heightened sensitivity to thermal pain at both sites (Cascio et al, 2007). Fig. 2 shows differential activity evoked in somatosensory cortex by vibrotactile stimulation of neighboring body sites using fMRI, demonstrable in individual subjects.

Baranek G.T., L.D. Parham, and J.W. Bodfish (2005). Sensory and motor features in autism: Assessment and intervention. In F Vokmar, A Klin, R Paul (Eds.), Handbook of autism and pervasive developmental disorders: Vol 2. Assessment interventions and policy (3rd Edition, pp. 831-857). Hoboken, NJ: John Wiley & Sons.

Cascio, C., F. McGlone, S. Folger, V. Tannan, G. Baranek, K.A. Pelphrey, and G. Essick. Tactile Perception in Adults with Autism: a Multidimensional Psychophysical Study. J Autism Dev Disord. 2007; Apr 6.

Essick, G.K., A. James, F.P. McGlone. Psychophysical assessment of the affective components of non-painful touch. Neuroreport 1999; 10(10):2083-7.

top of page

Principal Investigator: John Gilmore, M.D.

Early Brain Development in Twins

Twin studies have been critical in determining the contributions of genetic and environmental factors to risk for neuropsychiatric disorders[28]. Imaging studies in twins discordant for neuropsychiatric disorders reveal discordant brain structure, indicating that discordant brain development in twins is associated with risk for neuropsychiatric disorders[29-32]. It is unclear when discordant brain structure in twins arises and what causes it, though it is likely the result of abnormal early brain development, as early brain development is increasingly being recognized as the basis for many neuropsychiatric disorders. Understanding the causes of discordant brain structure in twins will significantly improve our understanding of the origins of neuropsychiatric disorders. Recently, sophisticated statistical modeling has been combined with high-resolution MRI and image analysis to determine the contributions of genetic and environmental factors to brain structure in adults[30, 33, 34].

Fig 13. Left-right asymmetry (A) and gender differences (B) in neonatal lateral ventricle volume. The left ventricle was significantly larger than the right (p = 0.0051, paired t-test). Female newborns have larger lateral ventricles than males (p = 0.0019).

Fig. 14. There was a significant correlation between gestational age at MRI and FA in the splenium of the corpus callosum (r2 = 0.5102; p = 0.0091). There was a similar significant correlation in the genu (r2 = 0.3639; p = 0.0291; data not shown).

Studies of normal monozygotic (MZ) and dizygotic (DZ) twins indicate that the size of many structures in the adult brain are determined by both genetic and environmental factors, while the size of others, such as the lateral ventricle, is determined mainly by environment[34, 35]. The available studies indicate that prenatal and neonatal brain structure in MZ twins are much more discordant than in adults, indicating that neonatal brain structure is under less genetic control than adult brain structure[36]. This suggests that genetic programs act over time during postnatal brain development to make initially discordant brains more concordant. Very little is known about early brain development in humans and how genetic programs and environmental factors regulate age specific neurodevelopmental processes. This study aims to provide critical, unavailable information about early brain development and the contributions of genetic and environmental factors to discordant brain structure in twins. This knowledge will directly improve our understanding of the meaning of twin studies of neurodevelopmental and neuropsychiatric disorders. The specific aims for this study are outlined below.

Aim 1: To study prenatal and neonatal brain structure and determine the degree and timing of discordant brain structure in MZ and DZ twins.

Aim 2: To determine genetic and environmental contributions to variation in prenatal and neonatal brain structure.

Aim 3: To explore prenatal and perinatal environmental predictors of discordance of brain structure in MZ and DZ twins.

Aim 4: To determine genetic and environmental contributions to specific aspects of cognitive development in the first two years of life; to determine if discordant neonatal brain structures predict discordant neurodevelopmental outcome in twins at two years of age.

Preliminary Results

With 20 neonates and imaged using the 3T Allegra scanner, lateral ventricle volume was determined by manual segmentation using the T1-weighted MP-RAGE sequence. In addition, apparent diffusion coefficients (ADC) and fractional anisotropy (FA) were studied using a region of interest (ROI) approach (genu and splenium of corpus callosum, internal capsule; occipital and frontal white matter). There was asymmetry of the lateral ventricles with the left having a larger volume than the right (p = 0.0051, paired t-test; Fig. 13). Female neonates had larger lateral ventricle volumes than males (p = 0.0019; Fig 13). There were no significant gender differences in gestational age at birth (males 279 ± 9 vs. females 271 ± 11 days), birth weight (males 3396.2 ± 551.6 vs. females 3546.1 ± 203.3 grams), age at MRI (males 19 ± 10 vs. females 16 ± 6 days), or ICV (males 484.8 ± 37.8 vs. 470.76 ± 58.8 ml). There was a significant correlation between gestational age at MRI and FA in the genu (r2 = 0.3639; p = 0.0291) and the splenium of the corpus callosum (r2 = 0.5102; p = 0.0091, Fig. 14). There was not a significant correlation for FA or ADC and gestational age at MRI in any other ROI. Compared to adults, neonates had significantly lower FA and significantly higher ADC in all ROI’s[37].

top of page

Principal Investigator: John Gilmore, M.D.

Early Brain Development in One and Two Years Old

The first two years of life is the most dynamic and perhaps the most critical phase of postnatal brain development[38, 39]. Concurrent with the rapid pace of structural brain growth is an equally rapid development of a wide range of cognitive and motor functions[40, 41]. In spite of its importance for understanding normal brain development as well as the early origins of neurodevelopmental disorders such as schizophrenia and autism, our knowledge of human brain development in this crucial time period is minimal. While there has been growing interest in MRI studies of normal human brain development, most studies to date have not included the first years of life as it is difficult to scan unsedated neonates and infants[38, 42-44]. In addition, image analysis of brain MRIs is very challenging in the first years of life. We have been able to address these challenges and have developed a large prospective cohort of children who have already had neonatal MRIs. In addition, we have developed cutting edge, state-of-the-art image analysis techniques that provide powerful tools to study brain development in the first two years of life[45-48]. Therefore, this study proposes to delineate brain development in normal one and two year olds. Please note that the proposed studies here differ from the project listed above for the twins. The proposed studies only consider singleton rather than the twins. This study will provide critical, unavailable information about early brain development and will ultimately provide the basis for future studies of neurodevelopmental disorders in this age group through pursuit of the following Specific Aims:

Aim 1. To study the developmental regulation of gray matter in the first two years of life.

Aim 2. To study the myelination and development of the corpus callosum, the corticospinal tract, and other white matter tracts with high resolution DTI and structural imaging.

Aim 3. To develop a brain atlas for neonates, one and two year olds, and map age related changes.

Aim 4. To determine if there are gender differences in brain development over the first two years of life.

Aim 5. To study the shape of the lateral ventricle over the first two years of life.

Text Box: Fig. 15. ADC and FA in a seed point in the midline of the genu and splenium of the corpus callosum and in myelinated and unmyelinated portions of the corticospinal tract.

Preliminary results

Recent advances on DTI techniques have allowed us to assess the maturation of white matter in the developing brain[37, 48]. A direct comparison of ADC and FA between neonates and 1yr old children in different brain regions is shown in Fig. 15. Note that the FA and ADC properties of the fiber bundles differ in their baseline in the neonatal period, as well as in the developmental trajectories of in the first year of life. The fibers of the corticospinal tract in the brainstem (seeded in the posterior limb of the internal capsule) are myelinated at birth; this region shows much less change compared to the unmyelinated cortical portion of the corticospinal tract in the cortex, as well as the genu and splenium regions of the corpus callosum (also unmyelinated at birth). Thus, the maturation of diffusion properties in specific regions of white matter tracts appear to be consistent with known myelination patterns in the first years of life. This novel approach will be able to detect subtle regional differences in development of fiber tracts over the first two years of life.

top of page

Principal Investigator: John Gilmore, M.D.

Prospective Studies of the Pathogenesis of Schizophrenia

Schizophrenia is associated with subtle abnormalities of brain structure, including enlarged lateral ventricles, reduced cortical gray matter volumes, reduced hippocampal volumes, as well as abnormal diffusion properties in white matter[49-52]. While it has been hypothesized that these brain abnormalities arise during early brain development[53, 54], there has been little direct evidence to support this idea. In the first funding period of this Conte Center project, we developed the MRI acquisition and image analysis tools to study very early brain development children at high risk for schizophrenia - the offspring of women with schizophrenia. Our study to date indicates that compared to normal controls, the offspring of mothers with schizophrenia have reduced cortical gray matter on neonatal MRI, while the lateral ventricles are not enlarged compared to controls. There is a suggestion of altered white matter development as well[45]. This is the first concrete evidence that genetic risk for schizophrenia is associated with altered gray matter development around the time of birth. The perinatal and early postnatal period is one of the critical phases in the development of cortical connectivity - a time of rapid synapse growth – one that is a focus of this Conte Center. In addition, we have developed a large cohort of normal controls. In this normal cohort, we have found robust growth of gray matter compared to white matter, as well as cortical region specific differences in gray matter growth in the neonatal period, consistent with human and non-human primate studies of synapse development[45].

In this funding period, we propose to continue our study of early brain development in normal and high risk children, applying our novel image analysis methodologies to study gray and white matter development in an expanding cohort, and to study longitudinal brain developmental as we follow our cohort into mid childhood. Specific Aims include:

Aim 1. Children at high risk for schizophrenia will have structural differences in gray and white development matter in the first two years of life compared to controls

Aim 2. Children at high risk for schizophrenia will have abnormal gray and white matter development at ages 4 and 6 years.

Aim 3. Children at high risk for schizophrenia will have unique brain structural abnormalities compared to the offspring of women with bipolar illness.

Aim 4. Specific risk genes will regulate gray and white matter development in normal children.

Preliminary results:

The feasibility of the proposed studies has been demonstrated by the above two projects led by Dr. Gilmore and thus is not repeated here.

Representative Publications:

Gilmore JH, Jarskog LF, Vadlamudi S. Maternal poly I:C exposure during pregnancy regulates TNFα, BDNF, and NGF expression in neonatal brain and the maternal-fetal unit of the rat. J Neuroimmunology 2005: 159: 106-112.

Jarskog LF, Glantz LA, Gilmore JH, Lieberman JA. Apoptosis in the pathophysiology of schizophrenia. Prog Neuro-Psychopharm Biol Psychiatry 2005; 29:846-458.

Prastawa M, Gilmore JH, Lin W, Gerig G. Automatic segmentation of MRIs of the developing newborn brain. Med Image Anal 2005; 9: 457-466.

top of page

Principal Investigator: Kelly Giovanello, Ph.D.

Functional-Anatomic Correlates of Relational Memory in Aging and MCI

The central aim of this study is to elucidate the core cognitive processes and fundamental neural mechanisms that give rise to relational memory impairments in aging and MCI. Anatomically constrained fMRI is utilized to assess prefrontal cortex (PFC) and specific structures within the medial temporal lobe (MTL) activations during memory performance in three groups of subjects: young adult controls (YC), healthy older adults who have intact cognition (OC), and subjects who meet criteria for amnestic mild cognitive impairment (MCI)[82]. The proposed experiments incorporate cognitive paradigms that directly manipulate the contribution of strategic and binding processes to relational memory performance and then examine the effect on functional neural architecture. Importantly, we will also examine the relationship between changes in functional activity and structural alterations that occur in aging and MCI.

Aim 1: Quantify changes in the neural architecture of relational memory under conditions that enhance encoding in aging and MCI.

Aim 2: Determine whether variations in retrieval mechanism influence functional activation in aging and MCI

Aim 3: Determine whether the neural mechanisms underlying generation and binding of relational information are similar across encoding and retrieval in aging and MCI.

Preliminary results

Text Box: Fig 23 fMRI studies on older adults.
In a recent study, we tested the hypothesis that healthy older adults – individuals who exhibit behavioral declines in relational memory – would show reduced specificity of hippocampal activity. During study, participants simultaneously viewed two nouns and were instructed to covertly generate a sentence that related the two words. To match performance, older adults were given three exposures to the study list. During retrieval, functional MR images were acquired while participants performed one of two recognition tasks (i.e., relational and item). In the relational task, participants were asked to indicate whether the two words were previously seen together. In the item task, participants were asked to indicate whether both items of a pair were previously seen. Relational and item conditions were intermixed with an active control condition and modeled separately in an event-related analysis. There was no difference in behavioral accuracy between young and older adults in either item (young M=.69; old M=.66) or relational (young M=.79; old M=.79) memory. In young adults, bilateral hippocampal activity was modulated by the extent to which the retrieval task depended on relational processing. Older adults showed activity in right hippocampus only, with activity in this region equivalent for item and relational memory conditions (see Fig. 23).

Representative publications:

Giovanello, K.S., Schnyer, D., & Verfaellie, M. (2004). A critical role for the anterior hippocampus in relational memory: Evidence from an fMRI study comparing associative and item recognition. Hippocampus, 14, 5-8.

Giovanello, K. S., Verfaellie, M., Keane, M.M. (2003). Disproportionate deficit in associative recognition relative to item recognition in global amnesia. Cognitive, Affective, and Behavioral Neuroscience, 3, 186-194.

Dew, I.T.Z., Bayen, U.J., & Giovanello, K.S. (in press). Implicit relational memory in aging. To appear in Zeitschrift fuer Psychologie.

Kan, I. P., Giovanello, K. S., Schnyer, D. M., Makris, N., & Verfaellie, M. (in press). Role of the medial temporal lobes in relational memory: Neuropsychological evidence from a cued recognition paradigm. Neuropsychologia.

Giovanello, K.S., Keane, M. M., & Verfaellie, M. (2006). The contribution of familiarity to associative memory in amnesia. Neuropsychologia, 44, 1859-65.

top of page

Principal Investigator: Peter Gordon, Ph.D.

Adaptation to Unilateral Hearing Loss in Humans: Cortical and Perceptual Effects

The ways in which sensory cortex in adults adapts in response to sensory loss is an issue of both theoretical and practical importance. Increased understanding of cortical adaptation could provide insight into how processes of learning are constrained both by patterns of neural connectivity created during development and by the limitations on neural growth in mature animals. Increased understanding of cortical adaptation could also shed light on how higher-level neural processes influence the benefits and limitations of interventions that restore (or partially restore) lost sensory input. While a great deal has been learned about cortical adaptation in adult mammals[78-81], there are important gaps in the understanding of this process in humans, particularly with respect to links between cortical adaptation and perceptual abilities.

The project tests individuals with unilateral, sharply-sloping hearing loss and individuals with normal hearing, all of whom will participate in a series of fMRI and psychoacoustic studies that pursue two specific aims.

Aim 1: To test the hypothesis that cortical lateralization is reduced for the unimpaired ear of unilaterally-impaired individuals even for frequencies where audiometric sensitivity in the impaired ear is normal.

Aim 2: To test the hypothesis that there is a link between cortical adaptation to unilateral-hearing loss and perceptual processing at frequencies where the impaired and unimpaired ears have equivalent audiometric sensitivity.

Text Box: Fig 22. Activation patterns for a patient, LM, with unilateral hearing loss and for her control. STIMulation was unilateral at 500 Hz, a frequency where LM has normal hearing both ears.

Preliminary Result

Fig. 22 illustrates the results of our studies by showing the activation maps from a patient (LM) and a matched-control participant when listening to unilaterally-presented stimuli at 500 Hz. The control participant (bottom row), showed evidence of the expected pattern of much greater activity in the contralateral auditory cortex compared to the ipsilateral cortex. This same pattern was observed in LM when this tone was presented to the impaired ear (top row, left column). However, there was a marked increase in ipsilateral activity when these stimuli were presented to LM’s unimpaired ear (top row, right column). These results show an increase in ipsilateral cortical activation for the unimpaired ear at frequencies where hearing in the impaired ear was fully preserved as indicated by audiometric criteria. This increase in ipsilateral cortical activity for the normal ear involves a region of auditory cortex that typically would be more heavily devoted to the impaired ear, a pattern that suggests that the normal ear may in some ways be dominating the impaired ear.

Representative publications:

Gordon, P.C., & Moser, S. (2007). Insight into analogy: Evidence from eye movements. Visual Cognition, 15, 20-35.

Camblin, C.C., Ledoux, K., Boudewijn, M., Gordon, P.C., & Swaab, T.Y. (2007). Processing new and repeated names: Effects of coreference on repetition priming with speech and fast RSVPs. Brain Research, 1146, 172-184.

Lee, Y., Lee, H., Gordon, P.C. (2007). Linguistic complexity and information structure in Korean: Evidence from eye-tracking during reading. Cognition, 104, 495-534.

Ledoux, K., Gordon, P.C., Camblin, C.C., & Swaab, T.Y. (2007). Coreference and lexical repetition: Neural mechanisms of discourse integration. Memory & Cognition. 35, 801-815.

top of page

Principal Investigator: Heather Cody Hazlett, Ph.D.

Brain development in school age children with autism

In this study, subjects will include children with autism, ascertained from a longitudinal MRI study of brain development in individuals with autism. Subjects will have already had brain MRI scans at ages 2 and 4. The proposed study will allow us to continue to track brain development as these children mature. Specific aims are

Aim 1: To characterize the longitudinal development of brain structure in autistic individuals from 2 to 10 years of age.

Aim 2: To characterize the longitudinal development of white matter using DTI in autistic individuals from 2 to 10 years.

Aim 3: To examine the development of selected brain structures and associated behavioral features between 2 and 10 years.

Preliminary results

We find significant enlargement of amygdala (AMYG) volume from 2 to 4 in our current sample of children with autism, compared to both developmental delay (DD) and controls (TYP) (Fig 17). The pattern observed is autism > DD > TYP. Amygdala enlargement was significant even after controlling for total tissue volume (TTV) (p = .0002). Both right and left AMYG were significantly enlarged compared to the TYP controls, while only the right AMYG was significantly enlarged compared to the DD controls. This finding is consistent with reports of AMYG enlargement in 3-4 year olds[55] and 7-12 year olds[56] with ASD.

top of page

Principal Investigator: Heather Cody Hazlett, Ph.D.

A pilot MRI study of infants at high risk for autism

This study proposes to conduct a pilot study of brain development in high risk baby sibs from ages 12 to 24 months. We hypothesize that overgrowth in white matter cerebral cortical volume will be detectable across the 12-24 month interval (i.e., marked change from 12 to 24 month) in infant sibs meeting criteria for autism spectrum disorder at 24 month. The specific aims are:

Aim 1: To demonstrate the feasibility of conducting a longitudinal MRI/DTI study of infant siblings of autistic individuals at 12 and 24 months; and to obtain pilot MRI/DTI and behavioral data on this population of children.

Aim 2: To characterize the longitudinal development of brain structure (regions, tissues and selected substructures) in autistic individuals from 12 to 24 months of age.

Aim 3: To characterize the longitudinal development of white matter using DTI in autistic individuals from 12 to 24 months.

Aim 4: To explore the development of brain-behavior relationships between 12 and 24 months.

Preliminary results:

This study is currently underway and we have had the opportunity to collect behavioral and MRI data on infants at risk for autism as young as 6 months of age. To date we have successfully scanned 4 infants at age 6 months and 5 infants at age 12 months.

Representative Publications

Ross A, Hazlett HC, Garrett N, Wilkerson C, & Piven J. (2005). Moderate sedation for MRI in young children with autism. Pediatric Radiology, 35: 867-871.

Wassink, T., Hazlett, H.C., Epping, E.A., Arndt, S., Dager, S.R., Schellenberg, G.D., Dawson, G., and Piven, J. (in press). Cerebral cortical gray matter overgrowth in autism is associated with functional variation of the serotonin transporter gene. Archives of General Psychiatry.

Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, & Gerig G. (2006). User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage, 31(3): 1116-1128.

C.1.1 Major benefits offered by the requested system for pediatric studies

Allow Dr. Lin’s project to devise dedicated pediatric phased array coils using 64 rf channels
Reduce total data acquisition time using parallel imaging

As extensively discussed above, although the total scanner time which includes calming the subjects, setting up monitoring system and imaging, may not substantially decrease with the requested system, the decreased acquisition time will have substantially impacts on the ongoing pediatric studies which include

improving the success rate and thus directly facilitating the proposed pediatric studies
minimizing the number of subjects needed to be excluded because of motion artifacts
minimizing the waste of resources resulted from motion artifact
shortening anesthetic time for the patient population (autism)

Improving the image quality for DTI images using parallel imaging methods

minimizing geometric distortion to facilitate the investigation of white matter maturation of the entire brain
enhancing SNR through the reduction of TE

Audiocomfort package (p31, Siemens quote): Minimizing acoustic noise during gradient switching and thus facilitating pediatric subjects since they are scanned during sleep.
BLADE package (p36, Siemens quote): This package may further improve the image quality for the listed pediatric studies.
AutoAlign package (p37, Siemens quote): This package will improve studies where subjects will be scanned multiple times.
DTI related packages (p38-43, Siemens quote): These packages directly enhance our ability to acquire high quality DTI images as well as analyze DTI images on the system directly.

C.2. Structural imaging

This section will summarize projects which focus mainly on structural imaging. Similar to Section C.1, a summary section will be provided at the end of this section outlining major benefits gained by the requested system with the exception of the benefits which are project specific.

top of page

Principal Investigator: Xuemei Huang, M.D.

Functional Studies of Subtypes of Parkinson's Disease

Parkinson’s disease (PD) is a neurodegenerative disorder having characteristic loss of substantia nigra dopamine neurons and clinical manifestations that include varying degrees of tremor, muscle rigidity, bradykinesia, and postural instability[83]. Cognitive impairment (in domains of frontal lobe function) is often a feature associated with the disease[84]. Clinically, patients with tremor-predominant PD (PDT) appear to have different pathophysiology compared to akinetic-rigidity predominant PD (PDAR). The former have different long-term motor outcome (i.e., fewer motor fluctuations, fewer levodopa-induced dyskinesias, and slower rate of progression), as well as fewer cognitive deficits. A classic model emphasizes the role of the basal ganglia in modulating cortical function through basal ganglia-thalamocortical circuits that ultimately cause bradykinesia and rigidity due to excessive thalamic inhibition in the dopamine deficiency state[85-87]. This model, however, does not explain the resting tremor in PD patients. Cerebello-thalamocortical circuitry has been known to be involved in the modulation of motor function, and a role also is implied in tremorgenesis (e.g., there are oscillating neuronal loops; tremor can be abolished by impairing the cerebellum-receiving areas of the thalamus; etc.)[88, 89]. Despite these data, the role of the cerebellum in the pathophysiology and symptoms of PD is not clear.

A new working model is proposed to explain the differential clinical manifestation of PD subtypes that involves two segregated, but functionally related, parallel loops of the basal ganglia-thalamocortical and cerebello-thalamocortical circuits. The working hypothesis is that these subtypes of PD have differential involvement of the two major motor regulatory circuits that is caused by both a different extent and location of neurodegeneration of melanin-containing neurons. Specifically, the following aims are proposed.

Aim 1: Understanding the neurocircuitry involved in PD, its clinical subtypes and stages.
Aim 2: Exploring structural MRI as a marker for PD and its progression.
Aim 3: Understanding the role of ApoE and LDL-cholesterol in PD etiology and progression.

Fig. 24. fMRI studies of PD twins

Preliminary results

Representative functional t-maps for the twin subjects performing the externally guided task at three axial levels are shown in Fig. 24. The twins displayed similar neural activation patterns, except that the PD-twin displayed relatively lower activation in subcortical structures and relatively heightened activity in cerebellum and PreMC areas.

Representative publications:

Huang X, Chen H, Miller W, Mailman R, Woodard J, Chen P, Xiang D, Murrow R, and Wang, Y.-Z. and Poole C. 2007. Lower LDL-cholesterol associated with Parkinson’s disease: a case control study. Mov. Disord. 22:377-381. PMID: 17177184.
George MS, Molnar CE, Grenesko EL, Anderson B, Mu Q, Johnson K, Nahas Z, Knable M, Fernandes P, Juncos J, Huang X, Nichols DE, Mailman RB. 2007. A single 20 mg dose of dihydrexidine (DAR-0100), a full D1 dopamine agonist, is safe and tolerated in patients with schizophrenia. Schizophr Res. 93(1-3):42-50. PMID: 17467956.
Lewis MM, Slagle CG, Smith DB, Truong Y, Bai P, McKeown M, Mailman RB, Belger A, Huang X. 2007. Task specific influences of Parkinson’s disease on the striato-thalamo-cortical and cerebello-thalamo-cortical motor circuitries. Neurosci. 147(1):224-35. Epub 2007 May 17. PMID: 17499933.
Huang X, Lee YZ, McKeown, M, Gerig G, Gu H, Lin W, Lewis MM, Ford S,, Troster A, Weinberger DR, Styner M. 2007. Asymmetrical ventricular enlargement in Parkinson’s disease. Mov Disord. 22(11):1657-1660 PMID: 17588238
McKeown M, Li J, Huang X, Lewis, MM, Rhee, S, Truong YKN and Wang ZJ, 2007 Local Linear Discriminant Analysis (LLDA) for Group and Region of Interest (ROI)-based fMRI analysis. NeuroImage 37(3):855-65. PMID: 1762785
McKeown M.J, Li, J, Huang X, and Wang J. Local Linear Discriminant Analysis (LLDA) for inference of multisubject FMRI data. 2007. The 32nd International Conference on Acoustic, Speech, and Signal Processing (ICASSP). (In press).

top of page

Principal Investigator: Weili Lin, Ph.D.

Characterization of Normal Brain Development Using Parallel MRI

This is a Bioengineering Research Partnership (BRP) grant where our partners include Dr. Larry Wald at the MGH and Dr. Guido Gerig at Utah University. The ultimate goal of this proposal is to develop dedicated imaging hardware and software for imaging very young normal children without sedation; results will allow a detailed characterization of normal brain development. Specifically, this application consists of two major components: 1) technical development of the required hardware (dedicated multi-channel phased array coils) and software (imaging sequences, reconstruction methods, and image analysis tools) and 2) a longitudinal study of brain development over the first two years of life. The specific aims for this project are outlined below.

Aim 1: To develop and optimize multi-channel phased array coils for the utilization of parallel imaging techniques to image normal pediatric subjects ages 2 weeks to two years old.

Aim 2: To develop parallel imaging methods, optimize imaging parameters for pediatric brains, and develop a “cloverleaf” navigator-based motion correction scheme to substantially reduce total data acquisition time while obtaining high resolution and high quality brain images

Aim 3: To conduct a longitudinal study over the first two years of life in order to quantitatively assess how the proposed technological developments improve pediatric imaging without sedation as well as to quantitatively characterize normal brain development

Preliminary Results

To illustrate the difficulties of imaging normal pediatric subjects as well as our ability to conduct pediatric imaging studies, Table 1 summarizes the total subjects that we have imaged and the success rates from several past and currently ongoing studies led by Dr. Gilmore and Dr. Piven (projects to be discussed below). The failure rate reflects obtaining no useful images and the success rates are divided into obtaining images with slight or no motion from at least 1 or 2, and all 3 imaging sequences (a successful study), respectively. The total scan time for the three sequences is about 25 min. Since the number of subjects for the 6 mon and 4yr groups are small, we will focus on the neonates, 1 and 2 yr old groups only. Clearly, the neonate group exhibits the highest rate of completing the entire study (64.8%) while both the 1yr and 2yr groups have a relatively low success rate, ~35%. The major reason for the low success rate for both the 1 yr and 2 yr old groups is motion related artifacts, underscoring the importance of reducing data acquisition time. To this end, our partner, Dr. Larry Wald at MGH is developing dedicated 8-channel phased array coils for the 2wk, 6 months, 1 yr and 2yr old groups, separately. Head/brain shape models for the various age groups of children have been generated based on an atlas-building approach and subsequently used to design the dedicated coils. Concepts of generating the head shape models and the representative examples from different age groups are shown in Fig. 4 and 5, respectively. With this approach, it will allow us to carefully design the dedicated head-coils for different age group.

Owning to the limited number of receivers (8-channel) available on our Allegra system, the proposed study focuses on developing dedicated 8-channel phased array coils. However, it has been pointed out that the extent to which data acquisition time can be reduced (acceleration factor) using parallel imaging approach highly depends on the number of rf channels available[1, 4]. G factors as a function of acceleration for the 8, 12, and 32 channel coils were obtained by measuring the noise correlation matrix (using an image acquired with no RF excitation) and coil sensitivity profile in a head shaped phantom (Fig. 6). For example, with an 8-channel phased array coil, a four-fold acceleration will result in a SNR loss of 69% (1-1/3.2) while a 32-channel phased array coil will allow 5-fold acceleration with a similar SNR loss. With an even higher acceleration factor, the SNR advantage clearly favors phased array coils with a higher number of receivers (79% vs 95% signal loss between 32- and 8-channel for a 6x acceleration). Therefore, with the requested TIM Trio 3T system, it will allow us to design dedicated phased array head-coils up to 64 channels, further reducing data acquisition time and improving success rates of imaging pediatric subjects without sedation.

Our partner, Dr. Larry Wald has extensive experience on devising phased array coil[5-12]. A representative example of a 96 channel phased array coil devised by Dr. Wald is shown in Fig. 7, demonstrating the technical expertise of Dr. Wald. Upon successful funding, Dr. Wald will modify the design of the head coils to take full advantage of the 64 channels. Dr. Wald has also agreed to work with our new hardware faculty (to be recruited, please see letter of support from Dr. Mauro, Radiology Chair) to further develop additional coils (letter of support from Dr. Wald is attached).

Another aspect of this project is to quantitatively determine white matter development using DTI[13-15]. Fig. 8 shows a comparison of white matter at different ages. Notice the elevation of FA values and the increased number of white matter bundles as a function of age, underscoring the importance of assessing white matter growth in the developing brain. While DTI has been widely used recently, one of the major difficulties for DTI has been geometric distortion as well as the lack of signal-to-noise ratio. With the utilization of parallel imaging methods, it has been demonstrated that it is possible to reduce TE which in turn improves the SNR. In addition, geometric distortion can be minimized using parallel imaging method, allowing improved DTI quality particularly in the regions such as the frontal lobe and the brain stem[1, 16]. Therefore, parallel imaging offered by the requested system is likely to further improve our ability to vigorously assess white matter growth in the developing brain.

Representative Publications:

Gilmore J, Lin W, Prastawa, MW, Looney CB, Sampath Y, Vetsa K, Knickmeyer RC, Evans DD, Smith JK, Hamer RM, Lieberman, JA, Gerig G. Regional Gray Matter Growth, Sexual Dimorphism, and Cerebral Asymmetry in the Neonatal Brain. J of Neurosciences, 2007, 27:1255-1260.

Gilmore J, Lin W, Gerig G. Fetal and Neonatal Brain Development. Am J Psychiatry: 2006, 163:12.

Prastawa M, Gilmore JH, Lin W, Gerig G. Automatic segmentation of mr images of the developing newborn brain. Med Image Anal. 2005;9:457-466

JH Gilmore, G Zhai, K Wilber, JK Smith, W Lin, G Gerig. 3T magnetic resonance imaging of the brain in newborns. Psychi Res. Neuroimaging, 2004, 132: 81-85.

top of page

Principal Investigator: Jair Soares, M.D.

Fronto-Limbic brain of bipolar Children and Adolescence

Abnormalities in fronto-limbic brain (FLB) regions involved in mood regulation have been associated with bipolar disorder (BD) in neuroimaging studies on adults[59-61]. Did these FLB abnormalities develop as BD progressed? Or, did they precede it, or emerge early during illness development? The latter would suggest that (1) developmental abnormalities in brain maturational processes lead to FLB changes. If so, then (2) treatments that reduce or alleviate these brain abnormalities early in the disease process might reduce the severity of BD. To test these hypotheses, we will conduct a combined in vivo neuroimaging and clinical intervention trial in adolescents who have BD. We will examine the relationships between (a) FLB abnormalities, (b) BD symptoms, and (c) response to a mood-stabilizing medication (valproate).

Aim 1: Neurodevelopmental FLB characteristics of BD adolescents.

Aim 2: Effects of mood-stabilizing medication on FLB abnormalities in BD adolescents.

Preliminary results

Volumetric measurements of different anatomical regions using manual segmentation were obtained. These regions include subgenual prefrontal cortex, singulate cortex, medial temporal lobe and corpus callosum. Specifically, results demonstrate that no differences were observed in prefrontal cortex among age-matched, BD, unipolar and normal controls. In contrast, the gray matter volumes of right and left anterior and posterior cingulate cortices were measured in 39 healthy controls (37 ± 10 years, 14 women), 11 untreated BD patients (38 ± 11 years, 5 women), and 16 BD patients on lithium monotherapy (33 ± 11 years, 7 women). With ANCOVA (age, gender and ICV as covariates), we found that untreated BD patients had decreased left anterior cingulate volumes compared to healthy controls (2.4 ± 0.3 and 2.9 ± 0.6 ml, respectively, p=0.016) and lithium-treated patients (3.3 ± 0.5 ml, p=0.001). Furthermore, BD patients had significantly larger left amygdala (2.57±0.69 vs. 2.17±0.58ml, respectively, p=0.040), and a trend for significantly smaller left hippocampus (3.87 ± 0.57 vs. 4.10±0.57ml, respectively, p=0.059) compared to healthy controls. The remaining temporal structures did not differ between the two groups (ANCOVA, age, gender, and ICV as covariate, p>0.05). Finally, no differences were observed in the corpus callosum among the patient groups.

All of these results were obtained from a 1.5T GE scanner. Dr. Soares has recently relocated to UNC from University of Texas Health Science Center at San Antonio. With the 3T scanner at UNC-CH, it is anticipated that image quality will be superior to that obtained using a 1.5T scanner. We have established the imaging protocols for Dr. Soares’ studies and the patient studies are currently underway using the Allegra scanner.

top of page

Principal Investigator: Jair Soares, M.D.

Neuroanatomy of Treatment Response in Bipolar Depression

Bipolar depression, the depressive phase of bipolar disorder (BD), is a frequent cause of chronic impairment and a major therapeutic challenge. Patients with bipolar depression are often non-responsive to available treatment interventions. The behavioral and clinical manifestations of BD are complex and likely to be mediated via a network of interconnected neuronal brain circuits. In particular, the fronto-limbic brain (FLB) circuits that are involved in mood regulation could be affected by the mechanisms that mediate the pathophysiology of bipolar depression[59-61]. A current hypothesis is that regional abnormalities in FLB regions in mood disorder patients could reflect disturbances in neuronal survival (or resilience)[62-64]. Such FLB abnormalities would generate the symptoms of depression, as well as the cycling seen in BD patients[60, 65]. Mood stabilizing medications such as lithium and specific anticonvulsants could have therapeutic effects via mechanisms that interfere with cellular resilience and confer neuroprotection, and perhaps reverse or alleviate focal brain abnormalities in FLB regions, if they have not advanced beyond a critical apoptotic stage[66-68].

This study will allow us to investigate, for the first-time, the contribution of FLB abnormalities to the pathophysiology of bipolar depression and mechanisms involved in mood stabilization. The specific aims of our proposed study are to test the hypotheses that bipolar depression in BD type I patients arises from impairment in neuronal survival in FLB circuits and that treatment response to mood stabilizing medications may involve amelioration or reversal of such abnormalities.

Fig. 19. Morphological findings in adult BP type I males: surface rendering of BD type I males left/right representative lateral ventricles (I-VI) and representative brain surface (VII-VIII) with Z-scores>2.5. The detected differences are color coded to indicate inward (blue) and outward (yellow) deformations needed to match the healthy male controls. The left lateral ventricle in BD type I was detected to be larger in over 69% of its surface, especially on the sup-mid, sup-post, inf-mid. and inf-post regions, views: I-III. The right BD type I lateral ventricle was detected to be larger on over

Preliminary results

We examined potential differences of brain surface and lateral ventricle shapes between adult BD type I subjects (n=32) and matched healthy controls (n=32). Although we detected no significant brain surface differences between BD type I patients and healthy controls in the entire sample, we found a slight enlargement of the left lateral ventricle in the superior-posterior region in male BD type I patients. Comparing female patients and female controls, there were no significant differences in either brain surface or ventricular shape. Over 69% of the left ventricular surface in male BD type I patients was identified by being significantly larger than that in male controls, being most significant in the superior and superior-posterior regions. The right lateral ventricle was also observed to be significantly enlarged over 32% of its surface. In regard to regional brain shape differences, our results detected an area that corresponds to the left dorsolateral prefrontal cortex that was significantly smaller in BD type I patients compared to controls (Fig. 19). Our findings of smaller left dorsolateral PFC in male BD type I patients further suggests that this structure plays a role in the pathophysiology of BD. This is the first time this structure has been shown to be anatomically abnormal in vivo in BD type I patients and the results are consistent with findings from functional imaging studies [59].

We examined correlations between amygdala volumes and NAA levels in dorsolateral PFC and cingulate volumes in a group of 47 BD individuals, in a preliminary study. This analysis yielded no significant relationship between right amygdala volumes and cingulate volumes and NAA levels in dorsolateral PFC, or left amygdala volumes and cingulate volumes, but demonstrated a significant inverse relationship between left amygdala volumes and NAA levels in dorsolateral PFC (Pearson correlation coefficient= -0.54, p=0.008; see Fig. 20). These findings are preliminary as they originate from a cross-sectional study, but point in the direction of our proposed hypothesis, suggesting that dysfunction of the PFC over time could possibly result on enlargement of amygdala.

Text Box: Fig. 20 Relation between NAA and left Amygdala volume.

Representative Publications:
Brambilla P, Macdonald AW 3rd, Sassi RB, Johnson MK, Mallinger AG, Carter CS, Soares JC.

Context processing performance in bipolar disorder patients. Bipolar Disord. 2007 May;9(3):230-7.

Monkul ES, Hatch JP, Nicoletti MA, Spence S, Brambilla P, Lacerda AL, Sassi RB, Mallinger AG, Keshavan MS, Soares JC. Fronto-limbic brain structures in suicidal and non-suicidal female patients with major depressive disorder. : Mol Psychiatry. 2007 Apr;12(4):360-6.

Najt P, Nicoletti M, Chen HH, Hatch JP, Caetano SC, Sassi RB, Axelson D, Brambilla P, Keshavan MS, Ryan ND, Birmaher B, Soares JC. Anatomical measurements of the orbitofrontal cortex in child and adolescent patients with bipolar disorder. Neurosci Lett. 2007 Feb 21;413(3):183-6.

Bearden CE, Thompson PM, Dalwani M, Hayashi KM, Lee AD, Nicoletti M, Trakhtenbroit M, Glahn DC, Brambilla P, Sassi RB, Mallinger AG, Frank E, Kupfer DJ, Soares JC. Greater Cortical Gray Matter Density in Lithium-Treated Patients with Bipolar Disorder. Biol Psychiatry. 2007 Jan 18; [Epub ahead of print]

C.3. Functional MRI studies

In the context of fMRI studies, the major advantages offered by the requested system are to minimize geometrical distortion and to improve spatial resolution through the utilization of parallel imaging. The former advantage will improve the image quality, particularly in regions such as the frontal lobe and the brain stem which have been difficult to image due to geometrical distortion using the current system. The latter advantage will better discern small regions of activation. Detailed information for each of the projects is given below. Similar to Section C.1, a summary section will be provided at the end of this section, outlining the benefits to be gained by the requested system with the exception of the benefits which are project specific.

top of page