shelf life and reduced susceptibility to disease are specific to anthocyanin hyper-accumulation, as there were no observed differences between control and Del/Ros1 fruits in firmness, cuticle thickness or cell wall composition at the full ripe stage. During over-ripening, Del/Ros1 fruits expressed significantly lower levels of multiple genes known to function in cell wall disassembly, such as polygalacturonase. B. cinerea is an important fruit pathogen and cell wall disassembly is integrally tied to its entry into the fruit [9]. The Del/Ros1 fruits have the double advantage of remaining firm for a longer period and accumulating substantially higher levels of water-soluble antioxidants. The oxidative burst is an important aspect of infection by necrotrophic pathogens that feed on dead tissue [10,11]. Clearly, the reduced susceptibility of fruits to postharvest diseases complements the reduction in fruit softening to enhance shelf life. The success of this strategy of increasing antioxidant capacity raises the possibility of extending the shelf life of other fruits. Much effort has gone into reducing losses of fruits to spoilage and the current work provides another avenue to that end. In summary, this biotechnological approach has the potential to increase the supply of a more nutritious food. References 1. Klee, H.J., and Giovannoni, J.J. (2011). Genetics and control of tomato fruit ripening and quality attributes. Annu. Rev. Genet. 45, 41–59. 2. Food and Agriculture Organization of the United Nations, Rome, 2011. Global Food Losses and Food Waste Extent, Causes and Prevention. 3. Khush, G., Lee, S., Cho, J.I., and Jeon, J.S. (2012). Biofortification of crops for reducing malnutrition. Plant Biotechnol. Rep. 6, 195–202. 4. Martin, C., Butelli, E., Petroni, K., and Tonelli, C. (2011). How can research on plants contribute to promoting human health? Plant Cell 23, 1685–1699. 5. Davis, D.R., Epp, M.D., and Riordan, H.D. (2004). Changes in USDA food composition data for 43 garden crops, 1950 to 1999. J. Am. Coll. Nutr. 23, 669–682. 6. Butelli, E., Titta, L., Giorgio, M., Mock, H.P., Matros, A., Peterek, S., Schijlen, E.G., Hall, R.D., Bovy, A.G., Luo, J., et al. (2008). Enrichment of tomato fruit with health-promoting anthocyanins by expression of select transcription factors. Nat. Biotechnol 26, 1301–1308. 7. Zhang, Y., Butelli, E., De Stefano, R., Schoonbeek, H., Magusin, A., Pagliarani, C., Wellner, N., Hill, L., Orzaez, D., Granell, A., et al. (2013). Anthocyanins double the shelf life of tomatoes by delaying overripening and reducing susceptibility to gray mold. Curr. Biol. 23, 1094–1100. 8. Jimenez, A., Creissen, G., Kular, B., Firmin, J., Robinson, S., Verhoeyen, M., and Mullineaux, P. (2002). Changes in oxidative processes and components of the antioxidant system during tomato fruit ripening. Planta 214, 751–758. 9. Cantu, D., Vicente, A.R., Greve, L.C., Dewey, F.M., Bennett, A.B., Labavitch, J.M., and Powell, A.L.T. (2008). The intersection between cell wall disassembly, ripening, and fruit susceptibility to Botrytis cinerea. Proc. Natl. Acad. Sci. USA 105, 859–864. 10. Glazebrook, J. (2005). Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annu. Rev. Phytopathol. 43, 205–227. 11. Segmuller, N., Kokkelink, L., Giesbert, S., Odinius, D., van Kan, J., and Tudzynski, P. (2008). NADPH oxidases are involved in differentiation and pathogenicity in Botrytis cinerea. Mol. Plant Microbe Interact. 21, 808–819. Horticultural Sciences, P.O. Box 110690, University of Florida, Gainesville FL 32611-0690, USA. E-mail: hjklee@ufl.edu http://dx.doi.org/10.1016/j.cub.2013.05.010 Neuronal Coding: The Value in Having an Average Voice Voices or faces that fall outside of the norm are the memorable ones. Recent human neuroimaging work, however, indicates that the average voice holds considerable currency for neuronal coding. The study also forges a bridge with the face recognition literature. Christopher I. Petkov and Quoc C. Vuong Being accused of having an average voice is not much of a compliment. Simon Cowell, who was a judge on the television show American Idol, has been known to tell contestants, ‘‘You look nice, but your voice is distinctly average, I’m afraid’’. Cowell might not appreciate how difficult it would be for him to identify, let alone judge, an extraordinary voice without a reference to the average or prototypical voice. A new human neuroimaging study [1], reported in this issue of Current Biology, provides strong evidence that a voice-sensitive region in the human brain relies on a representation of an average male and an average female voice. Neuronal Coding of Voice Identity The brains of humans [2] and other social animals [3] contain temporal lobe regions called ‘temporal voice areas’ (TVAs), which are particularly sensitive to sound acoustics associated with voices. It has been unclear, however, how these brain regions code for voice identity. An earlier notion of the neural coding of identity is that of a ‘grandmother cell’ [4]. In the extreme, single neurons encode single identities, which could ruin a family reunion if a grandmother’s cell became injured. Apart from interesting examples, such as a ‘Jenifer Aniston neuron’ in the brain of a patient who was familiar with the actress [5], even the researchers who identified these neurons conclude that there is little support for the notion of a one-to-one mapping of neuron-to-identity [6]. Sorry, grandma. Another prominent view is that ‘objects’, including faces or voices, are represented as points in a high-dimensional space (Figure 1A) [1,7–10]. For voices, the axes of this space represent acoustic features. The distinctiveness between any two voices can be represented by their geometric distance in this space. That is, the more distinctive that two voices are, the further that they are from each other in this space. These distances have been shown to influence how people judge voice [10,11]. Thus, they could also influence how the brain encodes individual voices. If the brain has a representation of an acoustic feature space, what is the reference point within this space? That is, how are voices related to each other? One possibility is to compute Dispatch R521