Panoptic Neuroanatomy: Digital Microscopy of Whole Brains and Brain-Wide Circuit Mapping

vidual neuron or action potential; relatively few electrodes (compared to the total number of neurons) can be put into the brain at a time due to mechanical/physiological constraints, and the live brain of any reasonably sized adult vertebrate can be probed with optical methods only over short distances because of optical scatter in live-brain tissue. There is, however, one level of analysis at which studying the entire nervous system as a unit is possible: neuroanatomy. There is, in principle, no limitation to studying the entire ex vivo brain, while still being able to access microscopic detail down to the level of individual proteins or organelles. Addressing the integration problem in neuroscience is a primary reason for such whole-brain, multi-scale neuroanatomy. Just as genomes have provided an integrative framework for cellular/molecular biology, so can whole-brain anatomical datasets provide a unifying informational framework for neuroscience. Neuroanatomy has often been referred to in a derogatory sense as ‘largely descriptive’. In contrast, digitized whole-brain datasets promise an era of quantitative analysis, while also providing a geometrical/spatial informational framework addressing the problem of discordant nomenclature [Bohland et al., 2009a] that hinders the inPhilosophers have pointed to the prevalence of the ‘mereological fallacy’ in contemporary neuroscience (attributing the properties of the whole to a part, a.k.a. the blind men and the elephant fallacy) [Bennett and Hacker, 2003]. Attention is often focused on the details of individual brain systems. Specific sensory or motor pathways are well studied and, increasingly, modulatory systems are as well (reward, sleep). However, integrative study of the whole nervous system is rare, and the knowledge remains fragmented. Overarching models, in the tradition of Skinner or Pavlov, are also oversimplified and therefore unable to grapple with the full complexity of brains and behaviors. In addition to focusing on brain subsystems, there is an increasing focus on specific model organisms, which makes this problem worse [Manger et al., 2008]. A basic reason for the difficulty of integrative study is instrumental. Despite technical advances, we do not have the means of simultaneously measuring all relevant variables (e.g. the state or activity of all neurons, including subcompartments). This is unlikely to change, due to basic physics/ physiology limitations. Whole-brain imaging techniques (PET, fMRI and MEG/ EEG) have fundamental spatial and temporal resolutions far removed from the indiPublished online: June 14, 2013