Researchers from Max Planck Institute for Biological Cybernetics in Tübingen and Max Planck Institute for Medical Research in Heidelberg, both in Germany, have reviewed in vivo multiphoton microscopy of the brain. Until recently, only electrophysiology could detect individual action potentials, but high-resistance electrodes can damage neurons and whole-cell patch clamping can change the activity of the neuron under study. These limitations, and the fact that electrophysiology does not provide visual information in real time, left a need filled by multiphoton microscopy with calcium-sensitive dyes. Thus, the authors emphasize that only multiphoton microscopy can resolve the activity and structure of neurons with single-cell and single-action-potential accuracy. This activity can be recorded from subcellular structures and from populations of neurons in vivo. The approach can enable recording from all neurons within the local area irrespective of cell type or activity — something that has eluded electrophysiological recording approaches.To allow calcium-sensitive dyes to enter neurons, ester modification has been used. Once inside, enzymes in each neuron cleave the ester group, and the dye becomes able to report action-potential activity. More recently, mice have been genetically engineered to express calcium-sensitive fluorophores in the brain, which should allow certain neuronal subtypes to be genetically targeted with these encoded indicators.Although multiphoton microscopy provides important benefits, it has some limitations. For example, as with many imaging methods, it remains difficult to achieve high temporal resolution. The resolution can be improved by implementing various scanning approaches, such as resonant mirror scanning or acousto-optical deflection. Additionally, although the technique enables deeper imaging than other optical methods — in part because it overcomes scattering — it cannot reach below approximately 1 mm deep. The depth is limited primarily because of overexcitation near the tissue surface, which leads to a spurious signal from that region. The future holds exciting possibilities. Three-dimensional multiphoton microscopy is under development in several labs in Germany. Virtual reality already has been used with the technique to image freely behaving animals. Multiphoton endoscopy is possible, and MRI or PET could guide the endoscope. (Nature Reviews Neuroscience, March 2008, pp. 195-205).