Home > Positioning > Persons > Edelman

Gerald Edelman (1929–2014)

Edelman proposed that the brain works like an ecosystem undergoing selection, not like a computer executing instructions. Neural Darwinism — the theory of neuronal group selection (TNGS) — argues that the brain develops and learns through a Darwinian process: a vast repertoire of neuronal groups (populations of interconnected neurons) is generated during development, and experience selects among them — strengthening the connections of groups that are useful and allowing unused ones to weaken and disappear. The analogy is not loose: Edelman drew it from his own Nobel Prize-winning work on the immune system, where the same pattern operates — enormous diversity is generated first (millions of antibody variants), and encounter with antigens selects the variants that fit. The brain, on this account, is not programmed; it is selected. The concept of reentrant signalling — continuous, bidirectional, simultaneous communication between brain areas — provides the mechanism by which selected neuronal groups are coordinated into coherent perceptual and cognitive states.

Life

Born 1 July 1929 in Ozone Park, Queens, New York. Undergraduate at Ursinus College (BS, 1950). MD from the University of Pennsylvania (1954). PhD at Rockefeller University (1960). Nobel Prize in Physiology or Medicine (1972, shared with Rodney Porter) for determining the chemical structure of antibodies — showing that antibodies are proteins composed of two heavy chains and two light chains, with variable regions that determine antigen specificity.

After the Nobel, Edelman turned from immunology to neuroscience — a shift motivated by the conviction that the immune system’s selectional logic applies to the brain. Founded the Neurosciences Institute (1981), initially at Rockefeller University, then relocated to La Jolla, California (1993), as an independent research institute devoted to the biological basis of higher brain function. The institute operated until 2012. Edelman directed it throughout, pursuing the neural-Darwinism programme with a dedicated research group.

Edelman published extensively for both scientific and general audiences. Neural Darwinism: The Theory of Neuronal Group Selection (1987) is the technical statement; Bright Air, Brilliant Fire: On the Matter of the Mind (1992) is the popular account. Died 17 May 2014 in La Jolla, California.

Neural Darwinism

The theory of neuronal group selection operates through three mechanisms:

Developmental selection. During brain development, an enormous repertoire of neuronal groups is generated — not by a genetic blueprint that specifies individual connections but by a process of growth, branching, and competitive interaction among neurons. The result is a primary repertoire: a population of neuronal groups, each with a distinctive pattern of connectivity, produced by a developmental process that is partly genetic and partly stochastic. No two brains — even genetically identical twins — have the same detailed wiring. The variation is the raw material for selection.

Experiential selection. After birth, experience selects among the neuronal groups in the primary repertoire. Groups whose activity correlates with salient environmental events are strengthened (their synaptic connections are reinforced); groups whose activity does not correlate are weakened. The process produces a secondary repertoire: a population of neuronal groups that has been shaped by the organism’s particular experience. Learning, on this account, is not the inscription of information into a passive medium (as in the computer metaphor) but the differential strengthening of pre-existing connections — selection, not instruction.

Reentrant signalling. The mechanism that coordinates selected neuronal groups across different brain areas. Reentrant signals are not feedback loops (which are sequential: A sends to B, B sends back to A) but parallel, bidirectional, simultaneous exchanges between multiple brain areas. The visual cortex, the motor cortex, and the association areas are continuously signalling to each other, synchronising their activity in real time. Reentrant signalling, Edelman argued, is the neural basis of perceptual categorisation and, ultimately, of consciousness — it is how distributed neuronal groups bind their activity into a unified perceptual experience.

Where Edelman stops

The selectional framework is compelling at the level of general principle (the brain generates diversity and selects from it) but difficult to test at the level of specific predictions. Identifying individual neuronal groups, measuring their competitive dynamics, and tracking their selection through experience requires experimental techniques that were largely unavailable during Edelman’s career. The theory was criticised by computational neuroscientists for being too vague — for describing the general character of brain organisation without specifying the computational operations that neuronal groups perform. Francis Crick reportedly said that Neural Darwinism was “neural Edelmanism” — more a framework than a testable theory. Whether subsequent advances in neural recording (calcium imaging, optogenetics, connectomics) have vindicated the selectional framework or rendered it unnecessary is debated within neuroscience.

The consciousness programme — developed in A Universe of Consciousness (2000, with Giulio Tononi) — proposed that consciousness arises from the integration of information across reentrant cortical circuits with high “complexity” (a measure of simultaneous differentiation and integration). Tononi subsequently developed this into the integrated information theory (IIT), which has become one of the leading theoretical frameworks for consciousness research. Whether IIT is a development of Edelman’s programme or a departure from it — Tononi’s information-theoretic framework is more formal and less selectional than Edelman’s — is a question in the recent history of consciousness studies.

The immune-system analogy that grounds the entire programme is productive but has limits. In the immune system, the selecting agent (the antigen) is external and identifiable; in the brain, the selecting agent (experience) is multidimensional and context-dependent. The immune system selects among antibodies that are functionally independent; neuronal groups are interconnected and their boundaries are not sharply defined. Whether the selectional framework captures the brain’s actual dynamics or imposes an immunological metaphor on a system that works differently is the deepest question in the theory’s reception.

Key works

See also: Darwin · Kauffman · Holland · Complex Adaptive Systems · Autopoiesis · Darwinism