Reflections on the past two decades of neuroscience


Recently, the journal Nature Reviews Neuroscience marked its 20th anniversary since the publication of its first issue. To celebrate this milestone, the editors invited prominent figures in the field of Neuroscience to write a Viewpoint article on what developments they consider particularly interesting and what research directions they envision for the future.


The contributing authors - among them Prof. Poirazi -, span a wide range of expertise: from glial neurobiology and neurodegenerative diseases, to computational modeling and neural prosthetics, While the perspectives presented in the article vary greatly, common themes arise.


Advanced experimental techniques

First thing not going amiss is the driving force of modern technological advances in the evolution of experimental techniques. Methods, such as optogenetics and chemogenetics, in combination with multi-photon microscopy and high density probes, have enabled simultaneous manipulation and recording of large populations of neurons with unmatched temporal and spatial precision. Such experimental setups would have seemed totally out of reach just a decade ago. Now, they are a daily lab routine, allowing researchers to explore previously inaccessible areas of neuroscience.


Computing power: a game-changer

On par with advances in experimental techniques, computing power has seen equally unprecedented progress, enabling faster data processing and greater sophistication in lab software design. In particular, the use of machine learning techniques for image analysis and pattern recognition has revolutionized our ability to automatically detect and classify patterns in the data, leading us to insightful revelations that we would otherwise miss in the haystack. This trend will only continue, and it's already being applied in clinical context.


The power of modeling

Beyond the data processing front, computing power is rapidly transforming computational and theoretical neuroscience. Modelers can now simulate complex neural circuits at scale, opening up new avenues of investigation into brain function.

Prof. Poirazi, a dendrites’ aficionado, has witnessed first-hand this 20 years’ worth of progress. Predictions proposed by her and other modelers have now been experimentally confirmed, and modeling studies are now routinely used to test theories and guide experimental design.


Online synergy

Regarding future research, she envisions experimental setups that would integrate computational models in-situ. Imagine a multi-compartmental model of a PFC pyramidal neuron dictating the optogenetic manipulation of the equivalent cell type in a live, behaving animal, all the while recordings of its neural activity are being fed back to the model synchronously. Such “online” model-controlled experiments hold tremendous potential in revealing hidden aspects of neural computation.


For more details check out the online version of the Nature Reviews Neuroscience article