arno Arno Klein

Phone:    917-512-5627
Open Science Framework:    papers, posters, proposals, and presentations

Current research: mobile health and brain imaging (see contributions to science).
As Director of Innovative Technologies at the Child Mind Institute in Manhattan, I am building a sensors and wearables program to study mental illness and offer potential interventions. Previous positions include Director of Neuroimaging at Sage Bionetworks, Assistant Professor of Clinical Neuroimaging at Columbia University, and Information Synthesis Theorist at the Parsons Institute for Information Mapping.


CMI_logo 2016 - Director of Innovative Technologies
Child Mind Institute, Manhattan

- Building a sensors and wearables program
  to study mental illness and offer potential interventions
Sage_logo 2014 - 2016 Director of Neuroimaging
Principal Scientist of Systems Biology
Sage Bionetworks, Seattle, WA
- Mobile health research app development
- Feature extraction from mobile phone sensor data
- Open science contests and crowdsourced data analysis
Stony Brook 2012 - 2013 Research Assistant Professor
Department of Psychiatry and Behavioral Science
State University of New York (SUNY) at Stony Brook
- Research imaging biomarkers of depression and PTSD
Columbia 2007 - 2012 Assistant Professor of Clinical Neurobiology
Division of Molecular Imaging and Neuropathology
Department of Psychiatry
New York State Psychiatric Institute
Columbia University
- Research human brain image processing, registration, and labeling
PIIM 2004 - 2007 Information Synthesis Theorist and Program Analyst
Parsons Institute for Information Mapping
The New School, NY
- Complex data visualization and information visualization
- Construction of visualization ontologies
Columbia crown 2004 - 2005 Research Scientist
Department of Psychiatry
Columbia University, NY
- Detect biomarkers of disease in brain MRI data


Weill 1998 - 2004 Weill Medical College of Cornell University, NY
Functional MRI Laboratory, Memorial Sloan-Kettering Cancer Center
Functional MRI Research Center, Columbia University
Ph.D. in Neuroscience, May, 2004
Thesis: Automated brain labeling with Mindboggle
- Invented Mindboggle to automate anatomical labeling of human brain data
- Simple mindreading based on task-evoked fMRI activity
Caltech 1996 - 1998 California Institute of Technology
Computation and Neural Systems Program
- Brain tissue optics research (with 2-photon microscopy and uncaging)
- Biophysical computer modeling of light propagation
- Kung fu and Muay Thai kickboxing clubs
MIT 1994 - 1996 Massachusetts Institute of Technology
Spatial Imaging Group, MIT Media Laboratory
M.S. in Media Arts and Sciences, September 1996
Thesis: Dispersion Compensation for Reflection Holography
- Dispersion correction for holographic view stations and edge-lit holograms
- Computer graphics for the electronic-holography display
MIT 1991 - 1993 University of Michigan (Ann Arbor)
B.S. in Biopsychology, Perception and Cognition Studies, May 1993
- Research assistant in the Kellogg Eye Institute
- Independent computer-generated holographic stereogram research
Waseda 1990 - 1991 Waseda University (Tokyo, Japan)
Japanese studies, International Division
- Independent autostereoscopic holography research, Tama Art College
- Shourinji kempo club, tournaments
USC 1988 - 1990 University of Southern California (Los Angeles)
Resident Honors Program scholar
- Research assistant in Hedco Neurosciences
- Independent display holography projects
- Kali-silat and kung fu clubs
USC 1985 - 1988 Ferndale High School (Ferndale, MI)
- 4.0 GPA
- Skipped senior year to attend college
- Display holography in a basement lab
- Cared for the school's reptiles, fish, and electric eel
- Taekwondo, cross country, track

Teaching and Service

2007 - 2015 Invited lectures about brain imaging research at:

-   Yale
-   IBM
-   UCLA
-   Google
-   Rutgers
-   UPenn (2x)
-   Dorkbot (3x)
-   Stanford (2x)
-   Janelia Farm
-   Columbia (10+x)
-   Child Mind Institute
-   Max Planck Institute
-   University of Konstanz
-   Stony Brook University
-   University of Washington
-   Seattle Hacker houses (4x)
2012 - Associate Editor for Frontiers in Brain Imaging Methods
2009 - Reviewer for neuroscience- and sensor-related journals including:

-   NeuroImage
-   Neuroinformatics
-   Human Brain Mapping
-   Brain Structure and Function
-   IEEE Trans. on Medical Imaging
-   IEEE Trans. on Biomedical Engineering
-   IEEE J. of Biomedical & Health Informatics
-   International Journal of Biomedical Imaging
-   Journal of Child and Adolescent Psychopharmacology
-   Psychiatry Research, Psychiatry Research: Neuroimaging
2009 - Member of the International Neuroinformatics Coordinating Facility's
Neuroimaging Task Force
2012 - 2013 Lecturer for Stony Brook University Medical Center's imaging seminar
2008 - 2010 Lecturer and guest lecturer for brain imaging courses, Columbia University
2005 - 2007 Lecturer on data visualization and visualization ontologies
to academic and government audiences, including:

-   National Academy of Sciences
-   Office of the Director of National Intelligence
-   National Geospatial-Intelligence Agency
-   Port Authority of New York and New Jersey
-   Under Secretary of Defense at the Pentagon
-   ESRI and GEOINT conferences
2005 - 2007 The New School (New York City):
-   MFA thesis evaluator and guest lecturer for design classes
-   Designed an M.S. curriculum for the Office of the President
-   Media Curricular Subchair under the Provost's office
2004 Review panel member, National Science Foundation
1994 - 1996 Teaching assistant (holography laboratory instructor), Media Laboratory, MA
1993 Algebra instructor, Washtenaw Community College, Ann Arbor, MI
1990 - 1991 English school instructor, Tokyo, Japan
1990 Graduate school mentor for holography projects, USC, CA


2015 - 2016 NCANDA-USA Consortium: Data Analysis Center
BD2K supplement
Role: Subcontract (15%)
Goal: Develop software to advance the integration and harmonization
of derived features and shape measures from NCANDA data
(National Consortium on Alcohol and Neurodevelopment in Adolescence).
2010 - 2014 Biological Predictors Software Supplement
NIH U01 grant supplement 3U01MH092250-03S1: $548,996
Role: Co-Investigator (90%); P.I.: Ramin V. Parsey
Goal: Extend feature extraction, identification, and shape analysis
algorithms in the Mindboggle software. Test prognostic accuracy of
feature-based biomarkers using EMBARC data (NIH U01 MH074813).
2012 - 2013 Biological Predictors of Response to Antidepressants
NIH U01 grant MH074813: $450,082
Role: Co-Investigator (25%); P.I.: Ramin V. Parsey
Goal: Reduce the trial and error associated with finding an effective
antidepressant by using data from pre-treatment quantification of 5-HT1A
receptors and 5-HTT to guide antidepressant treatment selection.
2009 - 2012 Mindboggling Shape Analysis and Identification
NIH R01 grant MH084029: $959,557
Role: Principal Investigator (80%)
Goal: Develop open source Mindboggle software to automatically extract
and identify brain features from MR images, label brain regions, and
measure the shapes of the features and regions.
2012 Ellora Documentation Project
Funded by the government of India.
Role: Co-Director
Goal: Develop and use a web application to geolocate the thousands of
sculptures and images of the Ellora cave temples of India with respect
to temple ground plans. Use this geolocation information to create
walkthroughs of the temples in the website
2006 Photodocumentation of the Ellora Cave Temples in India
Mellon Foundation grant
Role: Co-Director
Goal: Create the world's first comprehensive photodocumentation
of the sculptures and images of the Buddhist, Hindu, and Jain cave temples
at Ellora. Annotate the thousands of photographs, and create a website to
curate this information.
1993 Holographic research
Independent research grant, University of Michigan, Ann Arbor, MI
Role: Principal Investigator
Goal: Create a multiplex holographic stereogram to display an animated,
computer-generated model of a chameleon in three dimensions.

Contributions to Science

Open science and mobile health research
I am a passionate proponent of open science, where researchers share data, code, resources and ideas, and where collective, collaborative endeavors are preferred over separate silos of independent research. At Sage Bionetworks, I have spearheaded several open science initiatives, and have helped to coordinate open biomedical challenges such as the Alzheimer's Disease Big Data DREAM Challenge. To vastly scale up open medical research, I have been heavily engaged in mobile health research projects. I was the scientific lead on the mPower app for tracking symptom severity in Parkinson patients, one of the first research applications built on top of Apple's new open source ResearchKit platform. I built the mhealthx software pipeline for extracting features from sensor data from apps such as mPower, and designed visualizations to present mhealth data to patients, clinicians, and researchers. I draw inspiration from the MIT Media Lab's "demo or die" motto. I do not focus my efforts as much on publication as on execution, but I have written about re-envisioning the scientific review process and am interested in ways of overhauling how scientific data and knowledge are gathered, curated, and accessed.

Brain image analysis
Brain image morphometry is almost universally restricted to computing volumes and thicknesses for labeled gyri and subcortical structures. To better characterize the anatomy and shapes of brains, as P.I. of an R01 project (MH084029) I oversaw the construction of a new brain labeling protocol and the world's largest manually labeled set of brain images, labeled according to this more consistent and accurate protocol. One goal was to create tools for extracting different kinds of brain features and to characterize their shapes in more detailed ways, so I created and continue to be the main developer of the open source mindboggle software for automated brain feature extraction, labeling, and shape analysis. Features include anatomical regions (like gyri and subcortical regions), sulcal folds, and fundus curves. Shape measures include two types of depth, two types of curvature, volume, thickness, Zernike moments, Laplace-Beltrami spectra, etc. Using the mindboggle software, I recently completed the largest brain shape studies ever conducted, one as part of the Alzheimer's challenge mentioned above, for estimating cognitive state based on MR data, and two others as part of a large-scale characterization of normal brain shape variation (articles in preparation). I also helped develop a new method called "concurrence topology" for analyzing high-order relationships in temporal data (such as functional brain imaging data) using persistence homology, itself a new method from algebraic topology.

Large-scale brain image processing software evaluations
With the rapid growth in the number of brain image processing algorithms, it has become extremely difficult to know how they compare with each other or which to choose for a given study. This is exacerbated by the fact that the developers of these algorithms rarely conduct a rigorous evaluation against other methods. To rectify this, I determined that it was extremely important to the field of neuroscience to conduct a thorough evaluation of existing brain image processing algorithms. To this end, I published the largest registration and brain extraction evaluation studies ever conducted and have since participated in evaluation studies of other brain image processing steps as well.

Invertebrate sleep and learning
I have collaborated with my identical twin brother for years, and these studies have focused on how invertebrate sleep impacts a society, when and where bees sleep, and how it affects their ability to communicate. We are now completing a study of brain activity of olfactory learning during sleep in bees.

Since the time I built a basement optics lab with a friend in high school to create holograms, I have been making display holograms and conducting research on 3-dimensional display technologies, at the University of Southern California, the University of Michigan, and at the MIT Media Laboratory. For my thesis work at MIT, I created the deepest dispersion-controlled viewing stations, and the thinnest edge-lit holograms, both of which use a hologram to present specially controlled light sources (pre-distorted wavefronts or pre-dispersed light) to a second, display hologram for sharper and deeper images. Since then, I have formulated a general raytracing equation for holograms. Holography has influenced my later research in surprising ways.