Are neuroimaging researchers headed the Sheldon Cooper way?
The Human Connectome Project, The Brain Activity Map Project, The Human Brain Project, and the BRAIN are currently invested across the globe in better understanding brain anatomy and function. The ultimate goal of all these projects is to be able to reconstruct the human brain – have a definitive computer model or a database of every neuron and its cell type and the connections between each individual cell.
Is this feasible? Leading scientists gave their perspectives last year and each one of them was confident of success within probably the next decade. There was also a collection of other big names in the field (Full disclosure: Know/Worked for a few of them) who published their thoughts on the current state of the field of brain imaging and outlined ideas on possible achievable goals. If you are a glass-is-half-full person, the possibilities are endless.
While giving his views, Christof Koch, the Chief Scientific Officer at the Allen Institute of Brain Science, mentioned three goals over the next few years – mesoscale connective map of the mouse brain, cell type classification based on electrophysiology and genetic makeup, and finally the functional properties of all those classified neurons. In a recent issue of the Nature journal, a large group of scientists from Allen Institute published findings of the initial mouse connectome study.
Midway through reading this paper, I couldn’t help reflect on the amount of work/manpower it takes to generate this data. For a perspective, anyone knows how much time it took Santiago Ramon y Cajal to complete his drawings of the cortical column or maybe Korbinian Broadmann to outline areas in the cortex based on cytoarchitecture?
Mapping the mouse brain
In the paper from the Allen Institute, the authors detail the entire procedure for obtaining a mesoscale map of the mouse brain. Using an automated process for imaging and slicing the brain tissue, 140 sections of an entire mouse brain were imaged in 18.5 hours, generating ~750 GB of data to analyze. Data analysis involved two key steps – signal detection and image registration.
Not going into too much detail, the data went through steps of signal detection to separate signal from noise, square root transform to remove second-order effects, median and large kernel low-pass filtering to remove noise, developing candidate signal object sets using adaptive edge/line detection and then filtering them.
Since the images were acquired through an automated process, they were all well aligned and thus easily registered to the 3D Allen Reference Atlas model. The segmented and registered image was then combined and used to generate a connectivity matrix. It is fair to assume these took more than 18.5 months to optimize.
At this stage, the results of such an exercise are really an afterthought. Anyone in the field of neuroanatomy would know by now which two regions are connected and the nature of their connection. This paper is more a broad view of a mesoscale map of the mouse brain – 469 injection sites and 205 target sites in 12 major brain subdivisions.
But, if you are a glass-half-empty kind of person and followed the progress of connectomics research, you know that the only organism whose neuronal wiring diagram is connected entirely is the c-elegans’. The nematode worm has a total of 302 neurons and about 7000 synaptic connections between those neurons. Similar efforts are currently ongoing in the drosophila (fruit-fly) which has about 135,000 neurons.
Views on developing connectivity maps of other animal models and/or brain regions within an animal have been presented as part of the Brain Activity Map project. Based on comments from the scientists who are in-charge of all the brain research programs, the idea is to build on results from lower organisms and finally crack the neuronal code of the entire human brain.
I am a big fan of the Big Bang Theory- the TV show with four geeky physicists and their love interests (more particle physics than women). If you are not in the US, please check this list of channels that air the show internationally. Sheldon Cooper is the central character in the show – a young physicist with his share of idiosyncrasies – and works on String Theory.
I am not an expert on String Theory, but I do know two things about it: (i) it is a Theory of Everything, a mathematical framework that describes all fundamental forces and forms of matter and an attempt to reconcile General Relativity and Quantum Mechanics, and (ii) despite years of research and multiple String versions including an 11-dimensional M-Theory, nothing much has been achieved to reconcile anything. In the latest episode of the show, Cooper, who has invested nearly a couple of decades into further developing the theory is disappointed for no new breakthrough and quits to do other things.
Yeah, I know this is a show and drama is an integral part of comedy too. But are we as neuroimaging researchers also headed in the same direction? After all, our holy grail is in mapping the brain of the most imperfect mammalian species. We at best have a rough idea of the total neurons in our brain and mathematically the number of synapses possible.
There is no line of lab bred humans with similar cytoarchitectures. We, seven billion of us, are unique in our own ways. Our brains developed based on our experiences from even before our first breath out of the womb. We cannot possibly have seven billion connectomes, could we? Are we headed the Sheldon Cooper way?
Oh SW, Harris JA, Ng L, Winslow B, Cain N, Mihalas S, Wang Q, Lau C, Kuan L, Henry AM, Mortrud MT, Ouellette B, Nguyen TN, Sorensen SA, Slaughterbeck CR, Wakeman W, Li Y, Feng D, Ho A, Nicholas E, Hirokawa KE, Bohn P, Joines KM, Peng H, Hawrylycz MJ, Phillips JW, Hohmann JG, Wohnoutka P, Gerfen CR, Koch C, Bernard A, Dang C, Jones AR, & Zeng H (2014). A mesoscale connectome of the mouse brain. Nature, 508 (7495), 207-14 PMID: 24695228
brain, mapping, neuro imaging, future, sheldon cooper, string theory