Home email Computational Neuroscience

Computational Neuroscience


What is computational neuroscience? (XXXVII) Codes and processes

Two types of problems can be attributed to the idea of neural codes. Initially, while working on my critique of the neural coding metaphor, I focused mostly on the epistemic problem (the first two parts of the paper). The epistemic problem lies in the fact that when we claim that Y represents a neural code for X and that Y metaphorically decodes by the brain, it implies that Y provides information about X simply under being in a legal relation to X. This information is not available to anyone outside of the body, who can see both X AND Y and knows the correspondence. The organism wouldn’t know if X or X’ caused the neural activation Y if it only observed Y. The organism cannot see Y as a primary representation for X. It could be a secondary representation if the organism can observe that Y has a legal relationship with Z. We must then account for that primary representation which cannot be based upon an encoding. Several other authors have made similar criticisms, in particular Mark Bickhard.

The epistemic problem, or as I prefer to call it, is the “easy problem” with neural codes. Addressing it gives rise to alternative notions of information based on internal relations, such as O’Regan’s sensorimotor contingencies, Gibson’s invariance structure, and my subjective physics.

There is an even deeper problem. It is about substance vs. metaphysics and how time is conceived or disregarded in this instance. It is addressed in the third section of the neural code essay and my reply to comments (especially the third). It will be explained by comparing the neural code to the genetic code. There are some problematic aspects with the idea of a “genetic code”, but in its most unproblematic form, there is a lawful correspondence between triplets of nucleotides and amino acids, which we can call a code. There are two types, nucleotides or amino acids, which are a combination of some stable entities (molecules). The translation is a process by which nucleotides can be transformed into amino acids. A process is not a substance. It may involve certain substances, such as e.g. While enzymes are an example of a process, it is their activity that makes the process unique. The code refers only to a legal relationship between two substances. It does not include the process.

This analysis makes “neural code” look quite strange. The neural end of the code does not contain any substance. It is a measurement of neuron activity at a specific time, such as the number of spikes in a given time period. This number can then be interpreted as the output of a process, a stable entity that can further be manipulated or transformed by other processes. This is what it is for the experimenter who takes measurements and makes calculations. This view from the organism’s point of view is quite puzzling. The activity of neurons is a process and not the outcome of any other process. Neurons don’t produce stable entities such as amino acids that can be used in different processes. A spike is not stable. It is a timed event that occurs during neural interaction. It is inconsistent to treat the signatures of processes like substances.

Previous articleCar Computer Exchange
Next articleGeorgia Tech Computer Science


Please enter your comment!
Please enter your name here