This page describes Andre's collaboration with investigators in South Africa studying the effects of HIV on neural development in children.

Data Location

/cluster/gerenuk/user/andre_hiv/HIV_UCT_longitudinalstudy

Note: /cluster/gerenuk/user/andre_hiv/HIV_UCT is all of Andre's development work on the project before 2014. For the most part, you will always be working in the HIV_UCT_longitudinal directory.

Subdirectories

data_freesurfer_5.3- mprage scans processed with FS 5.3

data_freesurfer_5.3_fixed_putamen- Bruce modified mri_ca_label to stop overlabeling the putamen

data_freesurfer_6.0- mprage scans re-processed w/ FS 6.0 when released in Jan. 2017

data_orig- all scan data for all time points from South Africa

demographics- tables and spreadsheets with info on subjects

FromUCTMarch2016- new scan data Andre brought back from S. Africa in Mar. '16 on his yearly visit

FS_frozen_binaries- modified versions of v5.3 mri_ca_label and mri_ca_register binaries used for fixed putamen

mergetest- Allison M. testing how demographics files would be merged w/ command line tools. can probably be deleted.

NOTES- self-explanatory

path_files- text files listing paths to input data used for analyses

publications- ones written by Boston and S. Africa people

scripts- used for processing and analyzing data

stats- self-explanatory

Subjects

total = 145 - according to our spreadsheet

Subject ID # (<subjid>) = 1 - 3 digit code; 1 to 252

FreeSurfer subject directories:

<subjid>_<5/7/9> = cross-sectionally processed time point

<subjid> = subject base

<subjid>_<5/7/9>.long.<subjid>= longitudinally processed time point

Data Processing

Andre's primary role in the project is technical development for image acquisition. As such, we don't have any pre-defined role in data processing. We see it as really Ernesta's data and want to be respectful of that. We help analyze the data in whatever way they ask us to, while they also do their own analyses. So far, they have asked for our help with the two most FreeSurfer-related analyses:

Longitudinal Stream

Children participating in the study were scanned (ideally) at 5, 7 and 9 years old. Some children only have data from one or two timepoints. The FreeSurfer Longitudinal stream was used to generate the most accurate morphometric measures possible.

To do:

1. Edit (QA/QC) v6.0 recons
2. Re-run longitudinal stream w. edited recons
3. Statistical analyses w. accurate morphometric measures

Parameter Maps

Two multi-echo FLASH scans were collected during 5 year old scans to estimate T1 relaxation time and proton density (PD). This was done using the FreeSurfer tool mri_ms_fitparms. Allison and Andre compared these parameter values for T1, PD, T2* between healthy controls and children infected with HIV, using v5.3 recon-all output. These comparisons should be rerun once the 6.0 recons have been QA'ed.

To do:

1. Register newer, edited v6.0 aparc+aseg.mgz to parameter maps.
2. Redo statistical analysis comparing values across groups

People Involved

From Cape Town:

Ernesta Meintjes (joint-PI on the project)- [physicist in the Department of Biomedical Engineering at the University of Cape Town (UCT)

Barbara Laughton (also joint-PI)- [ bl2@sun.ac.za ] pediatrician at Stellenbosch University (sometimes abbreviated SUN) and Tygerberg Hospital (east of the city), where the children come for study.

Ken Mbugua- graduate student in biomedical engineering. He visited the Martinos Center for a couple of weeks two or three years ago and now he's analyzing the spectroscopy data and the 5 year old morphometry data.

Frances Robertson- Admins in Ernesta’s lab? set up data copying for us

Keri Woods- Admins in Ernesta’s lab?

Martha Holmes- demographics person (American. post-doc? did lot of analyses for spectroscopy. neuroscience person. well organized)

Emmanuel Nwosu - graduate student using HIV data for thesis/dissertation.

Mark Cotton- [ mcot@sun.ac.za ] professor at Stellenbosch University. on some of the abstracts.

From MGH:

Andre van der Kouwe- developed technical acquistion methods for un-sedated children, oversees MGH work for grant

Allison Moreau- RA working on the project. processed structural MRI data through FreeSurfer longitudinal stream pipeline and calcuated parameter maps from multi-echo FLASH data.

History

Our study is part of a larger study that was an international drug trial (CIPRA- Comprehensive International Program of Research on AIDS). South African piece and then Cape Town subset of that. CHER (Children w/ HIV early antiretroviral therapy)- local study name. piggybacked on for imaging. now can get demographics.

R21

Dates of funding: ?

The project was originally an R21 titled "Neuroimaging Technology for Pediatric Development Disorders in South Africa". Ernesta and Andre developed a set of techniques for imaging young children without sedation:

• use custom-designed motion-corrected MR pulse sequences
• can get statistical significance w/ fewer subjects, long analysis will increase power to see subtle differences between groups
• post-processing of DTI data for motion correction
• consistent on-line positioning on the scanner
• offline post-processing for distortion correction

This grant funded the study of the 5-year-old children for the HIV Project.

R01

Dates of funding: 7/1/2011 - 6/40/2017

This grant continued the work of the HIV project by providing funding to continue scanning the same (and new) children at 7 and 9 years old, using the techniques developed during the R21.

BACKGROUND AND SIGNIFICANCE

Since the advent of highly active Anti-retroviral Therapy (HAART), HIV has transitioned from a fatal disease to a chronic condition in which there may be ongoing damage to the central nervous system (CNS) especially in the developing brain of the young child. Because CNS penetration by ART is limited, the brain becomes a reservoir for the virus, and few drugs are available that impact these reservoirs. NeuroAIDS affects children differently and more dramatically than adults. Few studies of school-age HIV positive children exist and even fewer include neuroimaging. These studies are practically non- existent in the developing world. In this project we have the unique opportunity to study the effects of HIV and ART on the developing brain over a period of 4 years in a well-characterized cohort of young children using neuropsychological and behavioural assessments combined with advanced neuroimaging techniques in the region with the highest prevalence of HIV in the world, and the population most representative of the world pandemic.

RESEARCH DESIGN AND METHODS

• 110 HIV-infected children, 70 HIV-uninfected controls will be enrolled in study
  • followed since 7 weeks of age
• study duration = 5 years
• clinical evaluation and blood labs done every 6 months at Tygerberg Hospital for HIV-infected children
• controls will be clinically assessed annually to make sure still HIV-uninfected
• medical records available include nadir recorded CD4 levels and highest plasma viral load (PVL) since birth
• neuropsych assessments will be performed in child's native language- English, Afrikaans or Xhosa

STUDY OBJECTIVES

• Compare neuropsych, behavioral and neuroimaging findings of these groups:
  • 1) HIV-infected children - split into 3 treatment arms
    • a. who received current clinical care standard - delayed ART (Arm 1 of original CHER trial)
    • b. who received early ART followed by interruption at 1 year (Arm 2 of original CHER trial)
    • c. who received early ART followed by interruption at 2 years of age (Arm 3 of original CHER trial)
    • * difference between arms = when antiretroviral therapy began, based on previous study
  • 2) Children exposed to the virus but not infected
  • 3) Healthy control children (not exposed to or infected with HIV)
• To correlate cognitive and clinical development over time with anatomical and metabolic brain development, using neuroimaging in HIV-infected children.

The four aims of the proposed study are:

1. Clinical and comprehensive cognitive assessment of children at 7 and 9 years of age at KID CRU:

• 1.1. Clinical observations (principally: CD4 cell count, plasma viral load).
  1.2. Neuropsychological assessment (Griffiths Mental Development Scale and Beery-Buktenica Visual- Motor Integration score).
  

2. Image children at 7 and 9 years at CUBIC in a 1 hour session:

• 2.1. High resolution anatomical imaging (brain morphometry and tissue characterization: cortical thickness,

  cortical T1 values, cortical volumes, subcortical volumes).
  2.2. High resolution multi-direction diffusion weighted imaging (white matter characterization: brain structure fractional anisotropy, tractography).
  2.3. Single voxel spectroscopy (brain metabolism: absolute concentrations and ratios of N-acetylaspartate, choline, myo-inositol and creatine in peritrigonal white matter, midfrontal gray matter, basal ganglia).
  

3. Technical development of pediatric imaging techniques and longitudinal data analysis:

• 3.1. Further development and validation of motion corrected imaging sequences esp. diffusion/T2 imaging.
  3.2. Algorithms for data optimization for longitudinal analysis including Jacobian (deformation) analysis.
  

4. Capacity building through exchange of technology, students and technical training:

• 4.1. Cape Town student visits to Boston for training in image acquisition and sequence development,

  pediatric imaging and longitudinal data analysis for brain morphometry.
  4.2. Longitudinal brain morphometry analysis courses in Cape Town.
  4.3. Implementation of computational resource for brain image analysis at CHPC.
  

We will acquire three types of imaging data:

Structural imaging: We will collect multiecho MPRAGE (MEMPR) for cortical thickness estimation, cortical parcellation, and estimation of subcortical structure volumes using Freesurfer software. If time permits, we will collect two multiecho FLASH (MEF) scans for estimating T1 relaxation time and proton density (PD).

Spectroscopy: We will collect spectra in left peritrigonal white matter, midfrontal gray matter and basal ganglia, using a PRESS sequence with 2 mL voxels and 64 excitations analyzed with LCModel to estimate metabolite ratios and absolute concentrations (since Cr concentrations increase with HIV-associated gliosis).

.Diffusion imaging: We will collect a high-resolution (2 mm isotropic), 30 direction, whole brain diffusion tensor volume and analyze with TrackVis, DTIStudio, and the TBSS tool in FSL.*