Research software development

  • I help develop and maintain software and the IT stack for the International Brain Laboratory.
  • I develop Datoviz, an early C/C++ engine for GPU high-performance scientific visualization.
  • I'm part of the core development team of VisPy, a GPU-based scientific visualization library.
  • I maintain phy, a Python software for large-scale electrophysiological data analysis, visualization, and manual curation.
  • I've contributed to the development of Brian, a spiking neural network simulator in Python.

Neuroscience research

  • I've developed mathematical models of spiking neural networks to study the role of spike timing in neural computation (PhD dissertation).
  • I've worked on data analysis of neural electrical recordings in rodents to validate theoretical models to experimental data (PhD dissertation).

Teaching

Shaken Baby Syndrome

For the past few years, I've been interested in the science of the so-called shaken baby syndrome (SBS), also known as abusive head trauma (AHT). This is a fascinating academic topic at the intersection of science, medecine, criminal law, and sociology.

Most healthcare and law professionals involved with child protection assume that when an infant brought to the hospital is found with specific medical findings, namely subdural and retinal hemorrhage, the child has almost certainly been a victim of inflicted traumatic head injury. The sole observation of these findings almost always leads to a social investigation and criminal prosecution for child abuse (even child murder in the worst cases). Children are frequently put in foster care, and caregivers may be convicted and incarcerated, on the sole basis of an authoritative medical report, even when there is no other evidence of abuse, trauma, or violence.

I experienced this kind of diagnosis myself in 2016 when my 5-month-old son was briefly given to social services after he was found with subdural and retinal hemorrhage at the hospital. Fortunately, he recovered perfectly after neurosurgery, and the childminder, once prosecuted on the sole basis of this diagnosis, was cleared four years later. Medical experts eventually ruled that the actual cause of his symptoms was not child abuse, but a neurological condition called external hydrocephalus which can also cause subdural and retinal hemorrhage.

This personal experience led me to investigate this subject in depth from an academic viewpoint. As many other researchers and physicians before me, I discovered a lack of sound scientific foundations behind medical determinations of SBS/AHT. Although the discovery of multiple unexplained traumatic injuries in a premobile infant surely warrants an investigation, isolated subdural and retinal hemorrhage may result from dozens of medical conditions such as external hydrocephalus, blood clotting disorders, or even short falls.

In fact, even the causal link between impact-free traumatic shaking and isolated retino-dural hemorrhage has a weak scientific evidence base, as shown by a 2017 systematic review performed by the Swedish governmental health agency. A systematic diagnosis process relying exclusively on these two signs comes with a great risk of missing other medical explanations, notably due to cognitive biases among clinicians and medical experts.

In 2017, I co-founded a non profit organization, Adikia, which provides victims of such misdiagnoses with access to support groups. I'm also working with medical, scientific, and legal experts around the world to investigate the global extent of these medical errors and miscarriages of justice, to better understand the root causes of this possible societal failure, and to find ways to better differentiate non-traumatic from traumatic medical findings in young infants.

Some references: