A reflex is a decision-making system because it includes taking an action. A reflex entails a simple response to a stimulus based on a rule learned over long, evolutionary timescales. Although reflexes can change within a lifespan, those changes are either predetermined genetically (as animals change from children to adults) or entail only very simple learning processes (limited primarily to habituation and adaptation effects).
The reflex is the simplest decision-making system. It entails a simple rule-based reaction. In a sense, the thermostat is a reflex. When it is too hot, the thermostat turns on the air conditioning; when it is too cold, the thermostat turns on the heat. In animals, a reflex rule is prewired into the agent, presumably genetically. This means that we can imagine the rule as being learned over evolutionary timescales. Animals that had appropriate responses (pull your hand away from a burning painful stimulus) were more likely to survive and to have children that survived. But fundamentally, the key to a reflex is the simplicity of the rule and its prewired nature.1
The advantages of a reflex are (1) that it can respond very quickly and (2) because it has been learned over evolution, it is present from the beginning in an animal and the animal can react correctly the first time it encounters the stimulus. You don’t have to learn to pull your hand away from a hot stove. (You really should learn not to touch the hot stove in the first place. But that’s a different system.)
Reflexes are simple things—sensory neurons recognize the stimulus, and these neurons connect either directly or through one or more intermediate neurons to action-taking neurons that drive the muscles.2 With vertebrates, this generally takes place in the spinal cord because the faster the system can act, the better the reflex evolutionarily. Remember, reflexes are responding to stimuli already present. That is, your hand is being burned. The longer you wait, the more damage is being done to your tissues. Having the responding neurons be as close as possible to the stimulus allows the system to react quickly. Taking the message up to the brain takes time.
Myelinated nerve fibers carry action potentials at a speed of 150 meters per second.A Given that the distance from your hand to your head is about a meter, the journey from your hand to your head and back could take 14 milliseconds. Given that your fingers are being burned, that’s 14 milliseconds you don’t have. (Imagine how long it would take a signal to go from the foot to the brain of a giraffe or from the tail to the head of a blue whale.) Reflex responses therefore tend to be located in the spinal cord, which is closer to the action.
Reflexes could be simple connections between a sensory neuron and a motor neuron, or even a sensory neuron and the muscle itself. But, in practice, most reflexes have intermediate neurons between the sensor and the motor.4 An interesting question is why there is often an intermediate stage within the reflex, given that the time between detection and response is so precious. The presence of intermediate neurons provides additional stages at which higher, more complex decision-making systems can step in to change the responses.5 This is how we can prevent reflexes from acting if we want. Think, for example, of that scene in Lawrence of Arabia we talked about in Chapter 6. Normally, Lawrence would respond by reflexively dropping the match or shaking it out when the flame reaches his fingers, but with his indomitable will, Lawrence refuses to let his reflexes react and holds the match steady. The presence of intermediate neurons allows more controls for the other systems to manipulate, but it also increases the time it takes the decision to go from sensor to action. Some reflexes do not include intermediate neurons, while other reflexes have several stages of intermediate neurons.6
Classic examples of reflexes are the fleeing reflexes like the wind-sensors in insects and the crossover oculomotor turning neurons (visual [oculo-] to motor) in the goldfish.7 Insects such as crickets and cockroaches have a set of very small hairs on antennae on their tails. Each hair can be blown by the wind but can only move in a single direction, back and forth. This means that the amount that the hair moves depends on the direction of the wind—wind that is aligned with the hair will move it the most; wind that is off by 90 degrees will move it the least. These hairs (wind-sensors) project to four interneurons that represent the four cardinal directions (ahead, back, left, right). These interneurons project to the leg-muscle control structures and make the animal turn away from that wind. Why does the insect have this highly evolved wind-response reflex? Because it can flee from a stomping shoe in less than 20 milliseconds.
A similar system exists in goldfish connecting the recognition of visual signals on one side with a sudden turn to the other. A goldfish can identify that there is a looming signal (such as a goldfish-eating predator, like a diving bird) and start to turn away in less than 20 milliseconds. Goldfish can hit speeds of 22 body lengths per second within about 10 milliseconds of movement initiation. That’s incredibly fast! Presumably, this remarkable reflex evolved because the goldfish and the diving bird are in an arms race to see who is the fastest, the hungry bird or the fleeing goldfish.
Reflexes can, and, of course, do change within a single lifetime. (Animals have life-cycles—tadpoles and frogs presumably need very different reflexes to survive in the very different environments they inhabit.) But these changes are themselves prewired into development. Human babies, for example, have a foot reflex called the Babinski reflex, where the toes curl and splay in a specific pattern in response to stroking of the bottom of the foot. The reflex is entirely normal (and is used as a check for intactness of the spinal cord) in infants. With development of the corticospinal tract (connections from the cortex in the brain to the spinal cord), the reflex disappears. In adults, a Babinski reflex indicates spinal cord damage, decoupling the cortical input from the base of the spinal cord.8
The problem with reflexes is their inflexibility. Reflexes are learned on an evolutionary timescale; they are prewired, genetically, into the creature. Although reflexes do change in response to repeated stimuli, the mechanisms of these changes are not particularly complex, and the potential changes are very limited. These effects have been most famously studied in detail in the aplysia gill-withdrawal reflex.9 An aplysia is a kind of sea slug, with a broad, thin gill floating out of its back. (Imagine a snail with no shell and a pair of silk fans.) It uses this gill to breathe oxygen out of the water. But, of course, this gill is fragile and easily damaged. Imagine the aplysia in the water: the waves pick up and the aplysia pulls its gill in. But if the waves are going to be large for a long time and not get worse, then it needs to open out again. The aplysia can learn to ignore a repeated stimulus (that is, it can habituate). However, even these potential changes are prewired. For example, the gill-withdrawal reflex in the aplysia can increase its response to certain stimuli (that is, it can sensitize), but these changes are due to prewired connections between specific neurons.
Even simple prewired reactions and potential changes can produce extremely complex behavior if the changes interact with the environment appropriately. Examples of the complexity that can be built out of simple circuits without memory or explicit representation can be found in Valentino Braitenberg’s delightful book Vehicles. Nevertheless, this inflexibility means that in the competitive game of evolution, reflexes can get an animal only so far.
While evolution can move populations pretty quickly (on the order of a couple of generations),B reflexes are not being learned within a single lifetime.11 To learn to respond differently to different situations, one needs to turn to one of the other decision-making systems.
Because the reflex is so well understood, it is well described in many medical and neuroscience textbooks, such as
• Dale Purves et al. (2008) Neuroscience. Sunderland, MA: Sinauer Associates.
A thorough review of the remarkable escape reflexes can be found in
• Robert C. Eaton (1984) Neural Mechanisms of Startle Behavior New York: Springer-Verlag.
A delightful book showing how complex prewired reactions can be is
• Valentino Braitenberg (1986) Vehicles: Experiments in Synthetic Psychology Cambridge, MA: MIT Press.