Caseorganic: Created page with "===Definition=== Neurological computing is a loose term covering the several ways computation and the nervous system have been brought into direct relation. Three distinct programmes travel under similar names, and separating them prevents a good deal of confusion: * '''Computational neuroscience''' — using computation to model how nervous systems work. * '''Neuromorphic computing''' — building computers whose architecture imitates nervous systems. * '''Brain–comp..."

2026-08-25T13:08:51Z

Created page with "===Definition=== Neurological computing is a loose term covering the several ways computation and the nervous system have been brought into direct relation. Three distinct programmes travel under similar names, and separating them prevents a good deal of confusion: * '''Computational neuroscience''' — using computation to model how nervous systems work. * '''Neuromorphic computing''' — building computers whose architecture imitates nervous systems. * '''Brain–comp..."

New page

===Definition===
Neurological computing is a loose term covering the several ways computation and the nervous system have been brought into direct relation. Three distinct programmes travel under similar names, and separating them prevents a good deal of confusion:

* '''Computational neuroscience''' — using computation to model how nervous systems work.
* '''Neuromorphic computing''' — building computers whose architecture imitates nervous systems.
* '''Brain–computer interfaces''' — connecting a nervous system and a computer so that each can act on the other.

===Computational Neuroscience===
The founding result is Warren McCulloch and Walter Pitts's 1943 paper "A Logical Calculus of the Ideas Immanent in Nervous Activity," which showed that idealised neurons with threshold behaviour could compute any logical function — the first formal claim that the brain is a computing device and that a computing device could be a brain. Alan Hodgkin and Andrew Huxley's 1952 model of the squid giant axon gave the field its first quantitatively accurate account of how a neuron actually fires, and remains the basis of the discipline.

===Neuromorphic Computing===
Carver Mead coined "neuromorphic" at Caltech in the late 1980s for analogue VLSI circuits that imitate the operating principles of neural systems rather than their appearance. The motivating observation is an efficiency gap: a brain performs an enormous amount of computation on roughly twenty watts, while conventional digital hardware performing far less spends orders of magnitude more, largely because of the constant traffic between separated memory and processor.

Neuromorphic designs attack this by co-locating memory and computation, by operating on asynchronous spikes rather than clocked values, and by remaining largely idle. Working examples include IBM's TrueNorth, Intel's Loihi, the University of Manchester's SpiNNaker and Heidelberg's BrainScaleS. Event-based vision sensors, which report only the pixels that changed, are the same principle applied to input.

Neuromorphic computing should not be confused with artificial neural networks. The latter borrow a metaphor from neuroscience and run on entirely conventional hardware; the former change the hardware.

===Brain–Computer Interfaces===
The interface programme runs from Hans Berger's first human EEG recordings in 1924 through Jacques Vidal's coining of "brain–computer interface" in 1973 to contemporary work in both non-invasive form (EEG, fNIRS) and invasive form (intracortical microelectrode arrays such as the Utah array; cortical surface grids). The clinically significant applications are in restoring communication and movement to people with severe paralysis, and the deep brain stimulation systems now routinely implanted for Parkinson's disease are an established, if one-directional, neural interface.

The persistent obstacles are biological rather than computational: electrodes provoke a foreign-body response that degrades signal over months to years, bandwidth from the surface is low, and the mapping between recorded activity and intention has to be relearned as the tissue changes.

===Relevance to Cyborg Anthropology===
Each of the three programmes rests on a claim about what a brain is. Computational neuroscience treats it as an information processor; neuromorphic engineering treats its architecture as a design worth copying; the interface programme treats its signals as an input device. These are metaphors doing technical work, and — as N. Katherine Hayles argues in ''[[How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics|How We Became Posthuman]]'' — metaphors that have consequences for what gets built and for how people come to understand themselves.

===Related Reading===
* [[Brain Computer Interface]]
* [[Implant]]
* [[Ross Ashby]]
* [[Artificial Intelligence]]
* [[How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics]]

[[Category:Cyborg Anthropology]]
[[Category:Computing History]]

Neurological Computing - Revision history