SENSORiAL REDUNDANCY: what is the purpose of multiple pathways?
Every second we are bombarded with information from different sensory systems, starting from commonly known five senses to proprioception and internal chemical receptors. Even one modality is usually gets to the central analyser via several parallel pathways and is being processed in parallel. That means that the key feature of our sensory systems physiology is parallel processing.
parallel pathways advantages
As we can imagine parallel processing makes system more complicated. Is it really necessary? We could easily imagine that only one pathway per modality and even one sensory system would be enough for survival. So, what is the advantage of multiple sensory pathways?
The first obvious reason is safety. We can compare evolution with engineering, in both the principle is the same: minimal functional core and excessive control system. If something goes wrong and a pathway breaks there ought to be some other way to perform the task. If one sensory systems is disturbed, for example, animal looses eyesight, it always can rely on hearing and smell. Same principle goes for movement control .
The same applies to one modality. As we will discuss below different features of the same stimulus are acquired differently, so if you destroy different pathways you get different pathologies. Cats with major damage to the superior colliculus of the midbrain loose ability to identify movement direction, but can still recognise objects. And the opposite happens with visual cortex lesion . If these pathways weren’t quite independent, a cat would be completely blind in both cases.
Another reason is connected to the first one, but less intentional. We can explain some excessiveness as evolution side-effect. ‘Evolution is tinkering' wrote Francis Jacob in 1977 , and new structures appear on the basis of old ones, sometimes leading to repeats and strange design. Caenorabtitis elegans has redundant chemical sense , parrots have two song-systems . As with genes duplications and insertions duplication and build up upon old pathways can lead to new complex features. This hypothesis has been suggested for language acquisition in humans 
Also, there is no evident reason for contralateral sensory input, though the slight advantage is explainable with reason 1 – break proof. Unfortunately, it works only if in case of spinal injury, ipsilateral pathways are damaged but contralateral are still intact, whereas usually the trauma leads to more general damage. So, more reasonable explanation is an axial twist which occurred on some point of vertebrate evolution.
But the main reason I would like to highlight is the modality segregation. The more distinct features of the object or event we an animal can discriminate, the more precise and appropriate will be his response to it . The world complexity inevitably leads labour division; distinctive properties require particularly attuned receptors. And later these receptors cannot directly converge on the same neuron, the whole point would be completely lost.
As we have seen already, even via single sensory system we process information about quite composite stimuli. Let us briefly go through the examples.
Since Hubel and Wiesel ground-breaking discoveries visual system remains a frontier of sensory physiology. Basically, what was shown that starting from different types of retinal ganglion cells to specialised areas of visual sensory cortex separate features of visual stimulus is sent separately, converging to more complex feature detectors on each layer of processing. This lead to a breakthrough not only in visual perception research, but the sensory acquisition field as a whole .
More then that, the properties of retina, magnocellular and parvocellular cells, lateral geniculate nucleus and area V1 layout represent the most efficient from engineering point of view way to transmit spatial, temporal and colour information via a constrained diameter of nerve[.
Distinct populations of cochlear nucleus neurons send projections to various targets, such as olivary complex and reticular formation. Some of the pathways has established function, e.g. fusiform cells send information about stimulus localisation to inferior colliculus , some of them remain unclear. However, the very finding of specific neuron types for each pathways sheds light on the parallel pathways specificity in the brain.
A perfect example of specialisation are spinal cord parallel pathways. Spinothalamic tract carries pain and temperature information, crude touch. Posterior column-medial lemniscus pathway (PCML) fine touch, vibration and proprioceptive information. Two parts of PMCL, tract of Burdach and tract of Goll, are also distinguished by the depth of somatosensory sensation but the main difference between them is, of course, topology. The tracts get projections from two halves of the body — T6 and above, or T7 and below correspondently. But this is a place coding principle, which is a next reason for having separate parallel pathways.
Even for pain, which still remains an enigma for sensory physiology, several parallel ascending pathways (via lamina 1 in the dorsal horns of spinal cord) has been distinguished: crude pain, sharp pain, cold sensation and possibly itching.
Surprisingly, even old chemical senses —olfactory and gustatory systems — might have parallel processing in their core. Still, little is known about human sense of smell, but it has been confirmed for honey bees . Possible parallel pathways from taste buds are also discussed .
For the animal to survive it is important not only to know that something has happened, but the location and intensity of stimulus as well.
Despite the ‘all or nothing’ single neuron response our nervous systems solves this problem by using interval and place coding. And that is where parallel processing has another advantage.
Receptive fields correspond to stimulus location, distinct body part or another stimulus feature, the firing rate of the neurons encode the stimulus intensity.
It is also long known that hair cells topology in organ of Corti encodes different sound frequencies .
Interestingly, a school taught scheme of special taste regions on tongue is actually an old myth. In 1976 it was already shown that the differences in taste acquisition in different regions are insignificant .
And, the last but not least, as was already mentioned somatosensory system is a typical representation of the spatial representation. Primary cortex cortex and ventral posterior nucleus of thalamus are both somatotopically organized. Though the body representations not isomorphic, which is vividly depicted in a well-known homunculus model, but the projection area correlates with the receptors density in the matching body part.
While working at the same time, different pathways have different processing speed.
This also can lead to function specialisation, this time to the answer response. Of course sensory pathways need motor and humoral systems to send signals to, but the earlier this happens, the faster the reaction will be. That is why older structures like colliculi in the midbrain, striatum, substantia nigra and thalamus are essentially responsible for fast survival reflexes like shifting attention to sudden visual and auditory stimuli.
The differences between parallel processing speed is also a plausible explanation for such cognitive phenomena as priming, awareness, subliminal perception and even consciousness.
As we have seen, separate features of stimuli, its intensity and localisation are processed via separate pathways. Without communications these systems would be like blind men touching an elephant. That is why integration is crucial for seeing the whole picture and ultimately, for normal functioning. Too little integration sensory processing disorder, too much – synaesthesia, which is not so bad, but often has certain inconvenience.
Multisensory interactions take place between the same modality pathways (comparing different features), or completely different systems, e.g. audio-visual integration , and also with motor and other non-sensory circuits. Thalamic and cortical areas play a central role in sensory integration, but it also happens on lower layers.
Understanding parallel processing and sensory integration is not only one of the grand goals of human physiology, but also would have a huge impact in robotics and artificial intelligence fields.
I would like to conclude with rather philosophical thought. Whichever pathways and modalities we talk about, actually all the information is basically the same. Whether it is temperature, electromagnetic wave or chemical reaction, in the end it is an energy of atoms. So, our brain takes some properties of a vast atom-constructed world and using only most important for survival features makes a model of it.
And it is indeed more advantageous to divide the sensory input into parallel ways for processing, as we have seen already. It provides crash protection, function specialization and better attention and decision control. After all, renown strategic thinkers Sun Tzu and Caesar agreed upon «If enemy forces are united, separate them». And, at the same time, as every road leads to Rome, in the end parallel pathways need to be connected.
 E. D. Young, “Parallel processing in the nervous system: Evidence from sensory maps,” Proc. Natl. Acad. Sci., vol. 95, no. 3, pp. 933–934, Feb. 1998.
 B. M. Yu, “Neuroscience: Fault tolerance in the brain,” Nature, pp. 1–2, 2016.
 J. M. Sprague, “Extrageniculostriate Mechanisms Underlying Visually-Guided Orientation Behavior,” Prog. Brain Res., vol. 112, pp. 1–15, 1996.
 F. Jacob, “Evolution and tinkering.” 1977.
 P. T. McGrath, Y. Xu, M. Ailion, J. L. Garrison, R. A. Butcher, and C. I. Bargmann, “Parallel evolution of domesticated Caenorhabditis species targets pheromone receptor genes,” Nature, vol. 477, no. 7364, pp. 321–325, 2011.
 M. Chakraborty and E. D. Jarvis, “Brain evolution by brain pathway duplication.,” Philos. Trans. R. Soc. Lond. B. Biol. Sci., vol. 370, no. 1684, p. 20150056–, Dec. 2015.
 M. de Lussanet and J. Osse, “An ancestral axial twist explains the contralateral forebrain and the optic chiasm in vertebrates,” Anim. Biol., vol. 62, no. 2, pp. 193–216, 2012.
 J. A. Gottfried, “Neurobiology of Sensation and Reward.” CRC Press/Taylor & Francis, 2011.
 R. H. Wurtz, “Recounting the impact of Hubel and Wiesel.,” J. Physiol., vol. 587, no. Pt 12, pp. 2817–23, Jun. 2009.
 D. C. van Essen and C. H. Anderson, “Information Processing Strategies and Pathways in the Primate Visual System .,” Knowl. Creat. Diffus. Util., vol. 2nd, pp. 45–76, 1995.
 N. B. Cant and C. G. Benson, “Parallel auditory pathways: Projection patterns of the different neuronal populations in the dorsal and ventral cochlear nuclei,” Brain Res. Bull., vol. 60, no. 5–6, pp. 457–474, 2003.
 K. Inui, H. Okamoto, K. Miki, A. Gunji, and R. Kakigi, “Serial and parallel processing in the human auditory cortex: A magnetoencephalographic study,” Cereb. Cortex, vol. 16, no. 1, pp. 18–30, 2006.
 A. D. (Bud) Craig, “PAIN MECHANISMS : Labeled Lines Versus Convergence in Central Processing,” Annu. Rev. Neurosci., vol. 26, no. 1, pp. 1–30, 2003.
 W. Rössler and M. F. Brill, “Parallel processing in the honeybee olfactory pathway: Structure, function, and evolution,” J. Comp. Physiol. A Neuroethol. Sensory, Neural, Behav. Physiol., vol. 199, no. 11, pp. 981–996, 2013.
 S. D. Roper, “Parallel processing in mammalian taste buds?,” Physiol. Behav., vol. 97, no. 5, pp. 604–608, 2009.
 B. Fritzsch, I. Jahan, N. Pan, J. Kersigo, J. Duncan, and B. Kopecky, “Dissecting the molecular basis of organ of Corti development: Where are we now?,” Hear. Res., vol. 276, no. 1–2, pp. 16–26, Jun. 2011.
 V. B. Collings, L. Lindberg, and D. H. McBurney, “Spatial interactions of taste stimuli on the human tongue,” Percept. Psychophys., vol. 19, no. 1, pp. 69–71, Jan. 1976.
 M. Kiefer, U. Ansorge, J. D. Haynes, F. Hamker, U. Mattler, R. Verleger, and M. Niedeggen, “Neuro-cognitive mechanisms of conscious and unconscious visual perception: From a plethora of phenomena to general principles,” Adv. Cogn. Psychol., vol. 7, no. 1, pp. 55–67, 2011.
 P. Kaposvári, G. Csete, A. Bognár, P. Csibri, E. Tóth, N. Szabó, L. Vécsei, G. Sáry, and Z. Tamás Kincses, “Audio–visual integration through the parallel visual pathways,” Brain Res., vol. 1624, pp. 1–7, 2015.
 C. Cappe, E. M. Rouiller, and P. Barone, “Multisensory anatomical pathways,” Hear. Res., vol. 258, no. 1–2, pp. 28–36, 2009.