Taking three‐dimensional two‐photon excitation microscopy further: Encoding the light for decoding the brain

The recent work by Ducros et al. (2013) appears dou-bly inspired by Hadamard, in both a technical andglobal sense. The technical part, discussed later, takesadvantage of the well-known Hadamard transform.Globally, the work fits with Hadamard’s understand-ing of invention and discovery (Hadamard, 1954): itappears that a sense of beauty plays a central role inthe means of discovery. In the article by Ducros et al.context, motivation, elegance, and a sense of beautyare the elements of the scientific path toward under-standing the mysteries of the brain. It is a story andan adventure that has important milestones datingback to the work of Camillo Golgi and Ramon Y Cajal(Spacek, 1992; Golgi, 1873; Ramon y Cajal, 1904). Gol-gi’s approach introduced an elegant staining andembedding method for preserving neuronal spatialand temporal relationships of the cell-types andconnectivity-based organization of the brain (Spacek,1992; Golgi, 1873; Ramon y Cajal, 1904). Over the lastdecades, projects aimed at creating multimodal brainatlases (Roland and Zilles, 1994; Toga et al., 2006) andbrain models (Amunts et al., 2013) have been carriedout and coupled with this need to perform in vivo mini-mally invasive measurements for the understanding ofneural functions and activity at different temporal andspatial scales. The light microscope has gained a cen-tral role for mapping brain circuitry (Osten and Mar-grie, 2013; Alivisatos et al., 2013) and new technologyis in the spotlight. For almost two decades, two-photonexcitation (2PE) microscopy (Denk et al., 1990) hasbeen utilized within the neuroscience framework dueto its important abilities, namely: (a) three-dimensional (3D) optical sectioning, (b) reduction ofbackground fluorescence, (c) increased sample pene-tration and improved signal-to-noise ratio (SNR)despite the extremely small cross-section, and (d) sig-nificant reduction of overall photobleaching or photo-toxicity effects due to confinement of the 2PE event(Diaspro et al., 2006; Diaspro, 1999, 2004). Suchadvantages sometimes conflict with the amount of col-lectable photons which is especially penalizing in thecase of poor SNR and therefore demanding higherintensities and the fact that 2PE photobleaching andphototoxicity are proportional to the cube or fourthpower of the excitation intensity (Diaspro et al., 2006;Diaspro, 1999, 2004). However, the main problembehind the utilization of conventional 2PE scanningmicroscopy lies in the impossibility of permitting con-current acquisition of multiple signals within a milli-second time scale dynamics as needed for in vivocalcium variations revealing neuronal network activ-ities (Saggau, 2006). Now, encoding and decoding arethe powerful features added to 2PE microscopy byDucros et al. (2013). They have developed a cleverbinary amplitude modulation sequence, coupled tomultisite interrogation and point detection, to reveal“hidden” signals from the brain. Ducros’ approach, interms of signal treatment, recalls the so-called Fell-gett’s advantage for signals dominated by detectornoise (Hirschfeld, 1976; Fellgett, 1951, 1952): a SNRimprovement due to Hadamard multiplexed measure-ments rather than one-at-a-time multiplexing. Theeffect of multiplexing using the Hadamard transformcoding technique on the SNR arises from the fact thatone gets a N