Neuroscientists are mourning the loss of a beloved colleague and innovator. At the age of 41, Rahul Desikan died of amyotrophic lateral sclerosis (ALS)—a devastating neurodegenerative disease that he studied. Despite being struck with an aggressive form of the illness in early 2017, Desikan continued to publish about the genetic underpinnings of neurodegenerative disease, and to advocate for fundamental changes to clinical research. He worked until close to his passing, on July 14, 2019.

Rahul Desikan

“I have never in my life met someone like Rahul,” wrote Elizabeth Mormino of Stanford University. “The combination of sheer intelligence, intense work ethic, unfaltering devotion to family, openness with colleagues, and creativity towards all aspects of life is extremely unusual to find, all in a single person.”

In his career, Desikan bridged the fields of neuroimaging, biomarkers, and genomics. His discoveries not only advanced basic knowledge about neurodegenerative disease, but also equipped the field with tools that continue to be essential today. He led the development of the most widely used atlas of brain regions used in neuroimaging, and developed a polygenic hazard score (PHS) for Alzheimer’s disease that gauges age-specific disease risk based on dozens of genetic variations. Desikan was working on similar polygenic studies for ALS when he first started experiencing symptoms of that disease.

Alzforum featured Desikan’s work numerous times. As his disease progressed he continued to share information about his research and comment on others’ work, communicating with reporters.

Desikan was born in New Delhi, and immigrated to Queens, New York, as a young boy. He attended the Bronx High School of Science, and earned undergraduate degrees in biology, neuroscience, and the classics, and an M.D./Ph.D. at Boston University. It was as a graduate student in Ron Killiany’s lab that Desikan made perhaps his most lasting contribution to the field, the Desikan-Killiany Atlas (Desikan et al., 2006). Desikan distilled magnetic resonance imaging (MRI) data from 40 brains into an automated structural map of the cortex. Based on the natural topography of the brain’s gyral folds, the atlas delineated 34 regions of interest (ROI). When integrated into the brain mapping software FreeSurfer, the atlas became the most widely used regional brain map in the field of neuroimaging, noted Randy Buckner of Harvard University.

“The atlas linked traditional brain anatomy to the digital age of science, providing an anatomical lens to see signals in any of the many modalities now available to brain imaging,” noted Buckner. “It is fitting, and in his style, that Rahul delivered to us this otherwise invisible digital matrix in vibrant full color in a July 2006 issue of Neuroimage,” wrote Buckner, who recalled the young Desikan as “part brain-science graduate student and part gregarious, high-energy DJ on the Boston scene.”

Killiany noted that the atlas came at a time when the standard practice in the field was to draw regions of interest on MRI scans by hand. “Today, with the use of this atlas, investigators around the world are able to reproducibly identify brain regions on MRI scans not only to study the morphometry of these structures, but also to use these regions as a stepping-stone for more advanced analyses of brain features, such as connectivity.”

As a radiology resident at the University of California, San Diego, Desikan helped chart the atrophy of brain regions throughout the course of AD, and identified associations between brain atrophy, AD biomarkers, and genetic risk factors in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort (Nov 2011 news; Desikan et al., 2012; Apr 2012Jan 2014 news). Desikan began to study genetic risk at UCSD, where he parsed genomic datasets to uncover a previously hidden AD risk variant in MAPT, the gene that encodes tau (Mar 2015 news). 

“I was one of Rahul Desikan’s mentors, but I learned more from him than I ever taught him,” wrote Linda McEvoy of the University of California, San Diego. “Rahul’s research productivity during his radiology residency at UCSD was astounding. I was in awe of his ability to develop deep and lasting collaborations with so many experts in different facets of neurodegenerative diseases, synthesizing their input to inform his own unique approach.”

Desikan dove deeper into computational genomics at the University of California, San Francisco, where he moved in 2016. He headed studies that identified shared genetic variants between AD and multiple immune disorders, and elucidated a common genetic architecture between frontotemporal dementia and ALS (Apr 2016 newsApr 2018 news). 

Celeste Karch of Washington University in St. Louis, who co-led the 2018 ALS/FTD study with Desikan following his diagnosis, noted his openness and kindness. “These characteristics really influenced how he did science, where nearly all of his papers involved collaborations with groups around the world,” she wrote. “Rahul’s work leveraged big data to understand and better predict risk for a number of neurodegenerative diseases. He was constantly innovating and just getting started.”

In 2017, Desikan debuted a polygenic hazard score (PHS) for AD. By pooling genotypes at 31 single nucleotide polymorphisms (SNPs) associated with AD risk, Desikan and colleagues could estimate a person’s chances of developing AD at any given age (Mar 2017 news). The PHS correlated with the burden of Aβ and tau pathology, cognitive decline, and brain atrophy. His was not the only polygenic score, but it was distinctive in that it estimated when, not if, AD would strike. In February 2019, Desikan extended this research, linking the PHS to neuropathology and atrophy in a region-specific manner and drawing on the neuroimaging tools he had developed in graduate school. He reported that the PHS foreshadowed impending memory decline in cognitively normal people beyond standard Aβ-PET imaging alone (Feb 2019 news). 

When San Francisco-based Dash Genomics began marketing a commercialized version of Desikan’s PHS to the public, researchers expressed concern. Though not involved in the commercialization, and communicating only via eye movements on a specialized computer screen, Desikan told Alzforum that PHS must still undergo validation before being useful in predicting disease onset in individuals. He noted that polygenic assessments could be instrumental in selection for clinical trials, or to find out whether response to therapy is conditional on genetic risk in existing drug trials. 

That idea that people will respond differently to therapies depending on their unique palette of genetic risk factors drove Desikan to advocate for a personalized approach to clinical research in his final months of life. In The Washington Post in April 2018, he challenged the research community to change their approach: “Although every clinical trial for ALS has ‘failed,’ if you look carefully at the raw data you will find that a subset of patients who were treated showed slower disease progression,” he wrote. “Instead of lumping all patients into one group, why not focus on this group of responders so we can figure out the genetic, cellular and clinical characteristics of these individuals? What if we had all the data we need to find cures but didn’t have the right tools to see it?”

Buckner wrote to Alzforum that this will stay with him. “The conventional answer to this challenge is to believe there are disease subtypes and we will do better if we pick the right people for treatment,” Buckner wrote. “But I take Rahul’s challenge to be deeper: Can we build tools to not just see the average response but with enough precision to see that a single person or a handful are benefitting from a treatment? I will miss Rahul, and I will particularly miss the opportunity for him to bring the hidden signals in this complex landscape into full color for all of us to see.”—Jessica Shugart

Comments

  1. Rahul embodied the true spirit of a scientist. He was endlessly curious and always optimistic that he could contribute with careful data analyses and novel approaches. I knew Rahul for almost 20 years. Even as a student, he was already working on solutions to big problems with new tools and methods. Throughout his career, he continued to expand his skill set from imaging to genetics and beyond. He was always patient with me, explaining math that I couldn’t grasp easily, guiding me how to think differently with exploratory approaches to large data sets. I definitely learned more from Rahul than he ever learned from me. Rahul showed us all what courage truly means, both scientifically and personally. I hope he will remain an inspiration to young investigators for decades to come.

  2. Rahul was extremely brilliant, but he was also so open and kind. These characteristics really influenced how he did science, where nearly all of his papers involved collaborations with groups around the world. I loved getting his excited calls after we had run some analyses, “OK, is this crazy? This is what I’m thinking …” Rahul’s work leveraged big data to understand and better predict risk for a number of neurodegenerative diseases. He was constantly innovating and just getting started.

  3. Rahul was an amazing person. I first met him when he was working on the Desikan Atlas for FreeSurfer, which became a worldwide imaging standard. Subsequently we worked together when he used data from the Alzheimer’s Disease Neuroimaging Initiative to perform groundbreaking image analyses. Then he pioneered development of polygenic risk scores with ADNI genetics data. Here at UCSF, and in the ADNI community and around the world, Rahul will be missed.

  4. I knew RS Desikan for about 10 years before meeting Rahul in person. As a new FreeSurfer user during graduate school at UC Berkeley, I relied heavily on the Desikan cortical atlas. I still use these parcellations today in my lab at Stanford. Very frequently I would find myself reading a paper by Rahul, a consistent occurrence throughout all periods of my training. I arrived in Boston in 2011 for my postdoc after Rahul had already left the Martinos Center, but it was there that I heard more about him as a person. Most notably, I learned that he was a graduate student at the time he created the atlas I had been using for the previous five years. That is not what I had in mind! (I had an image of an older, established investigator, not someone my own age!) After I published my first polygenic risk paper in 2016, Rahul sent me an email to offer congratulations. He wrote, “My warmest congrats on your wonderful PRS paper in Neurology.” I was struck by the warmth of this email, not to mention how flattered I was that Rahul Desikan was taking the time to write to me! He was also beginning to explore aggregate genetic risk factors. We emailed about potential collaborations, and I remember vividly inquiring about meeting at the upcoming AAIC. He said he wouldn’t be going this year, since his wife was nearing the end of her pregnancy with their son. My respect for him continued to grow.

    When I started my position at Stanford in January of 2017, Rahul and I were neighbors since he was at UCSF. We finally set up a time to meet in person. It was around this time that the community learned about Rahul’s diagnosis. I had heard about this news from my many colleagues who had worked closely with Rahul, and it was very upsetting. I wasn’t sure if our meeting would still happen … I confirmed the location with him a few days ahead of time and he quickly confirmed. We met for coffee across the street from his lab. His speech was slightly slurred. We sat outside and talked about Boston. After a few minutes he shared with me his diagnosis, and told me it was important for him to tell me in person. We talked about his wife, his children, and his lab. We talked about science; we had a friendly quibble about mediation versus modification. We talked about the genetics of Alzheimer’s disease. We walked to his lab, and I meet his staff and trainees. We discussed a collaboration that resulted in a few publications in the years to follow.

    I have never in my life met someone like Rahul. The combination of sheer intelligence, intense work ethic, unfaltering devotion to family, openness with colleagues, and creativity towards all aspects of life is extremely unusual to find, all in a single person. In the face of illness riddled with uncertainty, Rahul continued to do what he loved, which was many things in work and non-work realms. Rahul is truly an inspiration, and I remain grateful for both learning from his research for over a decade, as well as learning from his view of love and purpose.

  5. I met Rahul when he was part brain-science graduate student and part gregarious, high-energy DJ on the Boston scene. His positive energy was infectious, which probably contributed to his style as a scientist. His best-known contribution was the Desikan-Killiany Atlas, the most widely used atlas of brain regions in the field of neuroimaging. That atlas’ adoption owes directly to its ability to link traditional brain anatomy to the digital age of science, providing an anatomical lens to see signals in any of the many modalities now available to brain imaging. It is fitting and in his style that Rahul delivered to us this otherwise invisible digital matrix in vibrant full color in a July 2006 issue of Neuroimage.

    Rahul’s lasting influence for me—and this will resonate with me for decades to come—is a challenge to the field that he laid down in the last months of his life, in a piece he published in The Washington Post at the end of April. Rahul reasoned that maybe clinical trials are failing so frequently not simply because we have failed to identify the cures, but rather because we are lumping all patients into one group when we should be focusing on the group of responders. He wrote: “What if we had all the data we need to find cures, but we didn’t have the right tools to see it?” That’s a deep and interesting thought—what if responders here or there are the key clues to advancement, and we’re just blind to their individual signals? What if the best we can hope for in some diseases is that a fraction of individuals will respond to any given therapy? The conventional answer to this challenge is to believe there are disease subtypes and we will do better if we pick the right people for treatment. But I take Rahul’s challenge to be deeper: Can we build tools to not just see the mean person shift, but with enough precision to see that a single person or a handful are benefitting from a treatment?

    I will miss Rahul, and I will particularly miss the opportunity for him to bring the hidden signals in this complex landscape into full color for all of us to see. I will remember Rahul and his great contributions in his too-short career, and most critically remember his forward-looking challenge to the field.

  6. The most inspiring talk I have ever attended took place in a small, cramped conference room at UCSF’s Parnassus Campus in the summer of 2018. It was a tour de force of Rahul Desikan’s body of work investigating genetic links between different neurodegenerative diseases. It was an amazing talk on its own merits, but when one considers that it was given entirely via a speech-assist device, and that the work had all taken place while Rahul was battling ALS, it was nothing short of remarkable. At the end, after he had finished taking multiple questions from the audience, I went up to Rahul to congratulate him. He was surrounded by his adoring and admiring family and colleagues. When he saw me he broke into his signature smile. It was the last time I ever saw Rahul, though we exchanged a few emails subsequently. It is a moment I will never forget and cherish forever.

  7. Rahul was a brilliant researcher who could contemplate large genetic databases and understand our genetic risk for neurodegenerative disorders. He had a vision for precision medicine that gave us a glimpse of the future. Simultaneously, with his passion and gentle soul, he had a unique way of uniting and inspiring scientists across the world. His life was cut short but his spirit touched so many.

  8. Rahul is one of the most amazing people I have ever had the opportunity to get to know in my career. He was clearly brilliant, with a capacity to grasp disparate concepts ranging from aspects of science, the etiology of disease, and realms of personal interest, such as music. When I first met him, he was struggling to learn how to control his immense cognitive abilities. Mental discipline was not one of his strengths when he entered his M.D./Ph.D. studies! His mind found it more rewarding to drift from coursework or board exams to concepts he deemed more important, such as ways to better understand how to use Magnetic Resonance Imaging to understand Alzheimer’s disease, music, or people who were close to him. At times this produced hurdles he then needed to cross. But there never seemed to be a hurdle he could not overcome in some fashion. Perhaps this is epitomized in how he tackled the final hurdle of the disease that ultimately claimed his life. He did so by not withdrawing from the things that were important to him, but rather by embracing them hard, leaving most of us to wonder toward the end how anyone could possibly have this much mental stamina.

    Rahul was not simply brilliant. From the time of his M.D./Ph.D. studies, I remember his respectfulness, courtesy, and quiet humility. He found it challenging to tell a more senior colleague or faculty member that they were wrong about something because he did not want to disrespect them. At the same time, he readily identified flaws in his own logic or methodological oversites. My colleagues remember fondly that he was always smiling even when the “errs of his ways” were being pointed out to him.

    Rahul’s contribution to science is nothing short of astounding, especially for someone who died so young. His publication record is stellar, starting with his early work in MRI, then moving into genetics. In collaboration with our colleagues at MGH, his creation of the Desikan-Killiany Atlas for use with FreeSurfer has exceeded any of our wildest imaginations. It came to maturity at a time where the standard practice in the field was to hand-draw regions of interest on MRI scans in order to identify various anatomical structures. Rahul firmly believed that using an automated tool to do this work would vastly improve the quality of our studies. Today, with the use of this atlas, investigators around the world are able to reproducibly identify brain regions on MRI scans, not only to study the morphometry of these structures but also to use these regions as a stepping-stone for more advanced analyses of brain features, such as connectivity. As a result, a wealth of studies being presented at major scientific conferences today cite Rahul’s name.

    I felt disbelief and then immense sadness when I learned of Rahul’s diagnosis, and even more sorrow when I learned of his passing this past July 14th. As mentors, our role is to advise and teach our students. In the long run, I’m not quite sure which one of us was actually the mentor and which one was the student—I am grateful for what he taught me.

    My thoughts and prayers go to his family and friends.

  9. I was one of Rahul Desikan’s mentors, but I learned more from him than I ever taught him. Rahul’s research productivity during his radiology residency at UCSD was astounding. He had unsurpassed energy, curiosity, and enthusiasm, bundled with a warm, engaging nature and keen sense of humor that made him a real joy to be around. I was in awe of his ability to develop deep and lasting collaborations with so many experts in different facets of neurodegenerative diseases, synthesizing their input to inform his own unique approach. I treasured our continued collaboration as he launched his career at UCSF and am devastated that it was cut short so tragically. There is no doubt he would have significantly advanced our understanding of neurodegenerative disorders had he not been stopped by ALS in such a cruel twist of fate. The world has lost so much. I am very fortunate for having been able to get know such a beautiful soul and will always strive to live up to the example he set, both before and after he was struck by this terrible disease. Rahul will never be forgotten.

  10. As a mentor, Rahul cared deeply about my scientific and academic career development. He was always thinking a few steps ahead for me and actively encouraged me to step out of my comfort zone. To me, he embodied the ideals of collaborative and interdisciplinary research.

    “Thanks, my friend,” “You too, dear friend” were frequent refrains of Rahul’s. When faced with new challenges I often find myself thinking, "What would Rahul have said or done?" I am grateful for this guiding beacon in my life. My dear friend Rahul, may you rest in peace.

  11. I never had the opportunity of meeting Rahul in person, but I obviously know about his work and we collaborated in a couple of manuscripts, so I can only talk about his research. However, even with that, it was clear that he was a different kind of researcher. He was extremely collaborative and his research went beyond identifying genetic regions associated with disease, to understanding what that data was telling about the biology of the disease. And, more importantly, how that data can be leveraged to help the patients with neurodegenerative disease. He never lost sight of the major goal, which is to help people with the disease.

  12. I met Rahul at a conference for early investigators several years back, and we all found it easy to become his colleague and friend. I was struck by his seemingly boundless energy, intelligence, and creativity. It's difficult for me to process the loss of such a gifted researcher and great person. His passing leaves a void not just for many of us personally but also for the future of our field. We can take some solace in the outstanding example Rahul provided in how to be a colleague and scientist.

  13. I met Rahul Desikan when we were both invited to attend the third annual Charleston Conference on Alzheimer's Disease in 2015 as Early Career Investigators. Literally within the first day of meeting, Rahul began running analyses to help me address research questions, and from then on, our collaboration grew rapidly and productively. Rahul was truly a “central hub” of collaborative science. Beyond that, he was a generous human being and a dear friend. Thank you, Rahul, for inspiring us all to be better people.

    Rahul Desikan (far right), Charleston, 2015.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. In Healthy Brains, Does Aβ Really Matter?
  2. Does Clusterin Interact With Aβ to Kick Off Neurodegeneration?
  3. Tau Gene Confirmed to Amp Up Alzheimer’s Risk and Neurodegeneration
  4. New Genetic Method Connects Immune Genes to Alzheimer’s
  5. Genetics Tie ALS into the Frontotemporal Dementia Spectrum
  6. Genetic Risk Score Combines AD GWAS Hits, Predicts Onset
  7. Multi-Gene Score Predicts Cognitive Decline Independently of Brain Imaging

Paper Citations

  1. . An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage. 2006 Jul 1;31(3):968-80. Epub 2006 Mar 10 PubMed.
  2. . Apolipoprotein E {varepsilon}4 Does Not Modulate Amyloid-β-Associated Neurodegeneration in Preclinical Alzheimer Disease. AJNR Am J Neuroradiol. 2012 Sep 13; PubMed.

External Citations

  1. Apr 2012

Further Reading

No Available Further Reading