Cognitive and perceptive functions are supported by highly overlapping neural networks distributed throughout the brain, and this phenomenon can be referred to as “neural reuse” (Anderson Reference Anderson2010; Reference Anderson2014; Pessoa Reference Pessoa2012). To use a metaphor, we might imagine the brain as a choir and neurons as the singers composing the choir; when the choir sings Song 1, some singers do not sing at all while others sing vigorously. “Active” singers represent the portion of the choir performing Song 1 (see Fig. 1a); in terms of the brain, those singers represent the neural network (parts of the brain) activated during a given cognitive or perceptive process (e.g., visual perception). When the choir sings Song 2, a slightly different, but highly overlapping portion of the singers might be active (see Fig. 1b). This exemplifies how different tasks (Song 1 and Song 2) recruit highly overlapping, but not identical, portions of the choir (see Fig. 1c); in the brain, a given cognitive or perceptive task (e.g., visual imagery) recruits a neural network highly overlapping, but not identical, to the one associated with visual perception (Ganis et al. Reference Ganis, Thompson and Kosslyn2004).

Figure 1. Singers of the choir singing (a) Song 1 (dark-grey); (b) Song 2 (dark-grey); (c) both Songs 1 and 2 (light-grey).

One of the hot topics for debate in experimental psychology and neuroscience is on the extent of specialisation or distribution of the brain functions (Anderson Reference Anderson2010; Driver & Noesselt Reference Driver and Noesselt2008; Liang et al. Reference Liang, Mouraux, Hu and Iannetti2013). Empirical evidence seems to favour the idea that the brain operates in a distributed, rather than specific, manner. For example, brain areas that were considered to be specialised for specific tasks (e.g., the fusiform face area for face recognition) have been found to be activated during performance of other tasks (e.g., recognition of cars; Gauthier et al. Reference Gauthier, Skudlarski, Gore and Anderson2000). Research has also shown that cognitive and perceptive processes usually involve networks of brain areas rather than one specific area; for example, a very specific process such as face recognition recruits a network of brain areas including the occipital, temporal, and frontal lobes rather than one specific portion of the brain (Goldstein Reference Goldstein2009). This distribution of brain functions is not only seen for the processing of faces, but also for other perceptual and cognitive functions (if not all; Van Dijk et al. Reference Van Dijk, Hedden, Venkataraman, Evans, Lazar and Buckner2010; Yeo et al. Reference Yeo, Krienen, Sepulcre, Sabuncu, Lashkari, Hollinshead, Roffman, Smoller, Zollei, Polimeni, Fischl, Liu and Buckner2011) such as perception (Takahashi et al. Reference Takahashi, Kansaku, Wada, Shibuya and Kitazawa2013; Uesaki & Ashida Reference Uesaki and Ashida2015), attention (Posner & Rothbart Reference Posner and Rothbart2007; Shulman et al. Reference Shulman, Ollinger, Akbudak, Conturo, Snyder, Petersen and Corbetta1999), memory (Alain et al. Reference Alain, Woods and Knight1998; Desgranges et al. Reference Desgranges, Baron and Eustache1998), language (Duffau Reference Duffau2008; Horwitz & Braun Reference Horwitz and Braun2004), spatial cognition (Burgess et al. Reference Burgess, Maguire, Spiers and O'Keefe2001; Vallar et al. Reference Vallar, Lobel, Galati, Berthoz, Pizzamiglio and Le Bihan1999), and body representation (Filippetti et al. Reference Filippetti, Lloyd-Fox, Longo, Farroni and Johnson2015; Longo et al. Reference Longo, Azañón and Haggard2010).

This phenomenon is not limited to the macroscopic level (i.e., brain areas), but has also been observed on the microscopic level. Some neurons have been found to respond to several types of stimuli, rather than to a specific type of stimulus, with different response patterns (e.g., firing frequencies). The difference in response patterns is the feature that distinguishes the stimuli represented by the neurons (Decharms & Zador Reference Decharms and Zador2000; Gerstner et al. Reference Gerstner, Kreiter, Markram and Herz1997). In a simplistic example, during the processing of Stimulus A, Neuron 1 fires at a high frequency, Neuron 2 at a low frequency, and Neuron 3 at a medium frequency. Whereas, during the processing of Stimulus B, Neuron 1 fires at a medium frequency, Neuron 2 at a high frequency, and Neuron 3 does not fire at all. To reuse the above-mentioned metaphor, the same singers might participate differently in performance of different songs.

The notion that brain functions are based on distributed and overlapping neural networks is convincingly supported by the findings that input from different sensory modalities activate distributed and overlapping networks of brain areas – namely, multisensory processing (Ricciardi et al. Reference Ricciardi, Bonino, Pellegrini and Pietrini2014; Stein & Stanford Reference Stein and Stanford2008; Stein et al. Reference Stein, Huneycutt and Meredith1988). Multisensory areas are portions of the brain processing input from different sensory modalities. In the last decades, an increasing number of multisensory areas have been identified (Amad et al. Reference Amad, Cachia, Gorwood, Pins, Delmaire, Rolland, Mondino, Thomas and Jardri2014; Gallese et al. Reference Gallese, Fadiga, Fogassi and Rizzolatti1996; Gobbelé et al. Reference Gobbelé, Schürmann, Forss, Juottonen, Buchner and Hari2003; Sereno & Huang Reference Sereno and Huang2006), suggesting that the brain is more engaged in multisensory processing than was initially believed. Classic multisensory (or associative) areas are activated by visual, auditory, and somatosensory input and consist of prefrontal (Fuster Reference Fuster1988; Öngür & Price Reference Öngür and Price2000), posterior parietal (Andersen et al. Reference Andersen, Essick and Siegel1985; Serino et al. Reference Serino, Canzoneri and Avenanti2011), and superior temporal (Beauchamp et al. Reference Beauchamp, Yasar, Frye and Ro2008; Bruce et al. Reference Bruce, Desimone and Gross1981) cortices. More recently, new multisensory areas have been identified, in the posterior frontal (Grafton et al. Reference Grafton, Fadiga, Arbib and Rizzolatti1997), temporoparietal (Matsuhashi et al. Reference Matsuhashi, Ikeda, Ohara, Matsumoto, Yamamoto, Takayama, Satowa, Begum, Usui, Nagamine, Mikuni, Takahashi, Miyamoto, Fukuyama and Shibasaki2004), and occipitotemporal (Beauchamp Reference Beauchamp2005) cortices. Additionally, some subcortical structures have been associated with multisensory processing, such as the superior colliculus (Jiang et al. Reference Jiang, Wallace, Jiang, Vaughan and Stein2001), hippocampus (Ravassard et al. Reference Ravassard, Kees, Willers, Ho, Aharoni, Cushman, Aghajan and Mehta2013), and amygdala (De Gelder & Bertelson Reference De Gelder and Bertelson2003). Widespread multisensory processing in the brain is responsible for the well-documented interaction of the senses during perceptual (Kawachi et al. Reference Kawachi, Grove and Sakurai2014; Vidal & Barrès Reference Vidal and Barrès2014) and cognitive tasks (Lawson et al. Reference Lawson, Boylan and Edwards2014; Pasqualotto et al. Reference Pasqualotto, Finucane and Newell2013a). Therefore, the hypothesis that largely overlapping brain networks underlie cognitive and perceptive functions is strongly supported by the findings that “overlapping” brain areas process input from different modalities.

Decisive evidence for the hypothesis that overlapping networks are the basis of brain functions comes from the surprising findings that multisensory processing also occurs in areas that were considered strictly unisensory, such as the primary visual (Borra & Rockland Reference Borra and Rockland2011; Zangaladze et al. Reference Zangaladze, Epstein, Grafton and Sathian1999) and primary auditory (Lakatos et al. Reference Lakatos, Chen, O'Connell, Mills and Schroeder2007; Murray et al. Reference Murray, Molholm, Michel, Heslenfeld, Ritter, Javitt, Schroeder and Foxe2005) cortices. Until recently, the notion that primary sensory cortices were concerned with processing input only from the corresponding sensory modalities was one of the core assumptions in neuroscience. However, this theory was challenged when studies with visually impaired participants surprisingly showed that the visual areas of these individuals were not “silent,” but they were active while visually impaired individuals performed a variety of tasks (Amedi et al. Reference Amedi, Raz, Pianka, Malach and Zohary2003; Poirier et al. Reference Poirier, Collignon, Scheiber, Renier, Vanlierde, Tranduy, Veraart and De Volder2006; Sadato et al. Reference Sadato, Pascual-Leone, Grafman, Ibañez, Deiber, Dold and Hallett1996). The activation of “visual” areas during nonvisual tasks performed by visually impaired individuals is a clear example of neural reuse, where a neural substrate deprived of its typical input is reused to process input from another modality (Guerreiro et al. Reference Guerreiro, Putzar and Röder2015; Iachini et al. Reference Iachini, Ruggiero and Ruotolo2014; Pasqualotto et al. Reference Pasqualotto, Lam and Proulx2013b; Saenz et al. Reference Saenz, Lewis, Huth, Fine and Koch2008). Additionally, sensory substitution, principally consisting of devices that convert visual information into its auditory or tactile equivalent (Bach-Y-Rita et al. Reference Bach-Y-Rita, Collins, Saunders, White and Scadden1969; Proulx et al. Reference Proulx, Ptito and Amedi2014), provides theoretical and practical insights into the ability of the brain (including the primary sensory cortices) to respond to environmental pressures (in this case sensory loss) by altering its functions.

Findings of studies with visually impaired participants demonstrate that even the primary visual cortex is involved in multisensory processing; but is this an effect of blindness? Studies wherein sighted adult participants underwent blindfolding showed that this is not the case. In fact, it is reported that participants who had been blindfolded for a few days subsequently exhibited activation of the primary visual cortex during performance of tactile tasks (Pascual-Leone & Hamilton Reference Pascual-Leone and Hamilton2001), and that this activation was necessary for successful completion of those tactile tasks (Kauffman et al. Reference Kauffman, Hamilton, Keenan, Warde and Pascual-Leone2000). Such a rapid effect of sensory deprivation on the brain function is incompatible with establishment of new brain connections and therefore suggests that multisensory processing “naturally” occurs in the primary sensory cortices (see also Hagen et al. Reference Hagen, Franzén, McGlone, Essick, Dancer and Pardo2002; Kayser et al. Reference Kayser, Petkov and Logothetis2008; Sathian & Zangaladze Reference Sathian and Zangaladze2002). Multisensory processing across distant parts of the brain is supported by preexisting brain connections that recent tracking techniques have started to uncover (Beer et al. Reference Beer, Plank and Greenlee2011; Reference Beer, Plank, Meyer and Greenlee2013; Kim et al. Reference Kim, Ducros, Carlson, Ronen, He, Ugurbil and Kim2006). In sum, mounting evidence indicates that distributed and overlapping neural networks encompassing both multisensory and “unisensory” (primary sensory) areas are underlying the brain functions, hence substantiating the idea that neural reuse is a ubiquitous phenomenon.

If highly overlapping parts of the brain are responsible for processing much of the information, how can we consciously undergo different experiences such as perceiving the smell of coffee or remembering the events of the past weekend? The answer is that highly overlapping parts of the brain are activated in a particular manner according to the content they process (Burgess et al. Reference Burgess, Maguire, Spiers and O'Keefe2001; Horwitz & Braun Reference Horwitz and Braun2004; Rolls & Tovee Reference Rolls and Tovee1995). Specific patterns of activation amongst overlapping neural population determine the types of neural processing and, ultimately, of the “mind content” (Shinkareva et al. Reference Shinkareva, Mason, Malave, Wang, Mitchell and Just2008). To recall the initial metaphor, different performance by each of the singers composing the same choir is responsible for the execution of a potentially unlimited number of songs; some up-beat, some down-beat, of different genres and styles.