Speaking from the White House in April, President Obama unveiled a major initiative to produce a brain activity map, which will transform our understanding of the brain and provide insight into various issues in neuroscience, from memory and Alzheimer’s disease to consciousness and creativity. This huge endeavour was greeted with widespread scepticism from neuroscientists, as no such map of the brain currently exists.
The 10-year project, Brain Research through Advancing Innovative Technologies, or BRAIN, will receive 110 million of public funding from the National Institute of Health, the National Science Foundation, and the Defence Advanced Research Projects Agency, for fiscal year 2014 alone. A similar amount has been promised by private sources.
When listing all of the future benefits of the initiative, Obama likened BRAIN to the Human Genome Project. This also took roughly a decade, generated research interest and funding, facilitated data-sharing and resulted in better technology. So far, so good. However, the Human Genome Project didn’t actually fulfil its main aim of improving the understanding of genetic disorders. In a similar vein, Obama announced that BRAIN will improve our understanding of psychiatric disorders like Schizophrenia and neurodegenerative conditions such as Parkinson’s disease – both of which are widely assumed to have genetic influences. If successful, our ability to explain, treat and even prevent these disorders would be vastly improved, but the limited progress made on such issues by the Human Genome Project suggests that the initiative is over-ambitious.
BRAIN isn’t the only initiative of its type – the Human Connectome Project, also funded by NIH, is studying long-distance, white matter connections; The Allen Brain Atlas is concerned with gene expression in the brain; and The Human Brain Project, aiming to produce a computer simulation of the brain, is scheduled to start on 1sOctober 2013. However, much of the criticism of BRAIN stems from the fact that no-one, including the researchers in charge, seems to know what the project will actually be doing, as it doesn’t currently have any clear goals or specific research aims. At the moment, the BRAIN Initiative is merely a modest pile of money with a fancy acronym.
One goal seems relatively established: better technology. A multitude of imaging and stimulation techniques are already at the disposal of neuroscientists; it’s possible to record the firing patterns of individual neurons (electrophysiology), monitor surface electrical activity (EEG), temporarily disrupt brain activity (TMS), map activity with radioactive tracers (PET) and notoriously, produce images showing increases in oxygen flow indicative of increased activity in certain brain areas (fMRI). So any new techniques emerging from the initiative are likely to be more precise, high-resolution versions of what is already around. This in itself would be valuable, because many current techniques are only suitable for use on animals and have limited resolution (detail level), affecting the tasks and functions they can suitably monitor, as shown in Figure 1 of ‘The Rebirth of Neuroscience…’, the graph of spatial resolution (the level of structure that can be measured or imaged using each technique) against temporal resolution (the amount of time over which activity can be measured). However, the kind of technology BRAIN is going to be developing can only be determined after a set of research aims have been established which specify the regions, levels of structure and substrates of interest. Aiming for ‘better technology’ alone is not sufficient.
In Obama’s statement, there was talk of trying to produce an activity map of the entire brain, like Human Genome Project did for the genome. The principles behind the two may appear very similar, but in reality producing a brain activity map which can be applied to any ‘typical’ human being is much more complicated. Firstly, the triplet code for proteins making up an allele’s coded-for protein had already been discovered; the Human Genome Project merely applied this principle on a much bigger scale. Neuroscience on the other hand has no such code, and although it is possible to associate areas of the brain with certain functions, the evidence is often weak or conflicting, and in much coarser detail than a locus of the genome. There is also a difference in the amount of variation complicating each project. Less than 2% of human DNA codes for phenotypes like eye colour and height, with only a limited possible amount of variation. The rest, so-called ‘non-coding DNA’, is thought to be involved in regulatory processes essential to sustaining life and therefore isn’t likely to differ greatly between individuals. Whilst subcortical brain structures also carry out regulatory functions, neurons in the cortex and their connections have a much greater potential for individual variation than DNA and the consequences of such variation that are much less severe. This must be accounted for in any general map of the brain. In other words, not all variation is pathological, making it more difficult to identify markers of pathology. Not only are these connections variable, they are also capable of changing according to an individual’s experience, being strengthened due to regular use, or otherwise pruned out. Accounting for neuroplasticity is thus a further complication of brain mapping projects. 10 years suddenly seems a rather tight deadline for a complete activity map.
Instead, BRAIN needs to focus on one kind of map. There are many possibilities; the map could be based on structure or function, and follow short- or long-distance connections, or a certain neurotransmitter. Ultimately, the map of choice should depend on some solid hypotheses since these will affect what the project can contribute to neuroscience, or in other words, how useful all that money will be.
Even with a smaller, more achievable mapping goal, there are still inference problems with these initiatives. Measuring activity whilst participants complete a task tapping certain abilities, as these projects propose to do, allows certain areas or connections to be associated with a particular function, but this evidence is correlational. It cannot be said that an area is causally necessary for that function, or what its role is. Correlations can only provide a very sketchy map; there is a need to move from correlation to causation to provide much stronger evidence.
Moreover, the collaborators behind BRAIN have asserted that by building a complete map it will be possible to ‘understand the brain’ and infer the nature and causes of neurodegenerative and psychiatric disorders. The Human Connectome Project also wants to first create a brain map and then draw inferences about the functions of each part. This approach assumes that the causes of neurological and psychiatric disorders can be found simply by looking at divergence from a brain map. At best, this can provide an insight into areas associated with the relevant impairments, thus allowing treatment of the symptoms, but the cause may remain unclear. Rather, several neuroscientists argue that it would be more prudent to examine the effects of such conditions and work backwards, using and contributing to, an appropriate map. Either way, a full understanding of such conditions is unlikely to be imminent, and perhaps BRAIN, like the Human Genome Project before it, will struggle to deliver anything concrete on the matter.
So, what can be expected from the BRAIN Initiative? A full activity map is highly unlikely, as is a complete understanding of disorders of the brain and mind. However, it may well generate more research funding and improved techniques are certainly welcome, whatever they prove to be. Nonetheless, it is questionable whether the outcomes of the initiative can match the media hype that surrounded its announcement. With the official goals of BRAIN due to be announced at the end of the year, it remains to be seen just how BRAIN will help or harm neuroscience’s endeavours and public image.