Study Materials

Decision-making, i.e., Principles of Behavioral Organization

Animals, at any given time in their life, have a repertoire of behaviors available to them. For instance, hiding, eating, reproducing. The ‘decision’ to choose one behavior over another is based on the immediate (and future) needs of the animal, and on the environment around them. Animals are ‘optimizers,’ making the best decision based on the information they have. Mechanistically, however, how do animals decide between and control their behaviors?

Behaviors are determined by interactions between genes and environment (e.g., learning, social status, etc.) across the lifespan. The behavioral repertoire also varies across the lifespan. Behaviors necessary early in the lifespan (e.g., suckling) are not necessary later, and vice versa (e.g., reproductive behavior). These long-timescale changes in behavior are mediated by hormonal systems (e.g., testosterone and the HVc in songbirds) and biological clocks (e.g., circannual and circadian rhythms). These two systems are not completely independent, for instance, the effects of biological clocks are usually mediated by hormones, and hormonal state can determine the response to a biological clock.

On a more immediate timescale, nervous systems control behavior. One theory says that behaviors are organized in systems called command centers – groups of interconnected neurons (neural networks) which turn behaviors on and off on a rapidly, e.g., the choice to fight or flee, or go get a drink of water. These networks, in addition to activating a response, inhibit other competing responses. It is not a good thing to try to drink at the same time as you are fleeing from an enemy. The behavioral choice depends on the pattern of neural input to the command centers (from sensory systems), and on the modulatory state of the animal (e.g., hormonal state, or motivational state, etc.).

As with many things in this class, it is improper to think of these principles as acting in isolation. Hormones can in fact control behavior on a very short (minute) timescale, and the output of the nervous system is modulated by hormonal state.

An example. The limb of a cockroach can participate in multiple behaviors. These include walking, righting and scratching. Each of these behaviors can be thought of as a command centers which provide output to the motor output system for the leg (in this case, a CPG, more on CPGs in a moment, but output systems are not necessarily a CPG!), which then controls the activity of muscles in the limb. The motor output system makes the limb move in a way which is appropriate to the intended behavior. Based on sensory inputs (e.g., is the roach upside down, thirsty, senses a mate or some mite crawling up its leg), the command center will make decisions which are relayed through motor output systems, and ultimately to the muscle making the animal’s limb move in an appropriate manner.

One example of a motor output system is a central pattern generator (CPG). These are neural networks that control rhythmic behaviors such as walking, swimming, or chewing (in the case of the lobster STG). The really cool thing about CPGs, is that it has been demonstrated that, depending on the pattern of neural input and the modulatory state of the network (which hormones or neuromodulators are present), the same EXACT neurons can produce a different output (i.e., different behaviors)! This is a principle that can most likely be applied to ALL motor output systems, and to the nervous system as a whole. Individual behaviors share much of the same neural circuitry, and how the circuitry behaves (and therefore the animal behaves) is dependent on hormonal, motivational, circadian, sensory, etc. etc. etc. inputs the circuit receives. Each behavior does not have its own independent circuit. Using the same circuits for many behaviors is a much more efficient way to build an animal than building many circuits which are rarely used.

Aside, can you also see where learning and plasticity would come in to alter the behavior of a given piece of circuitry?

Cladistic Analysis

This figure represents a cladogram or a phylogenetic tree. In other words, the evolutionary relationships between the species A, B and C, and their common ancestor D. A and C share the common trait x, and B has the trait y. It is most likely that the common ancestor D also shared trait x. This is because of the Law or Principle of Parsimony - if two species share a trait, it is most likely that their common ancestor shared the same trait, and that events of evolutionary change are rare, and so you should postulate the minimum number of these events. In this case, it is most parsimonious to say that x is ancestral (shared by D) and that y was derived, or evolved by B at point c – one step of evolutionary change. Less likely, but possible, is that y was ancestral (shared by D and B) and that x evolved or was derived twice independently by A and C at points b and d. This is a less parsimonious explanation for the evolution of traits x and y.

One way to test these two hypotheses is to do an outgroup analysis, i.e., find an ancestral line that split before D (e.g., ask whether species E has trait x or y). If E has trait x, then it is most parsimonious to say that F had the ancestral trait x (D also x), and then there would be only a single derivation of trait y at c. However, what happens if E has trait y? In this case, there are two equally possible hypotheses. First, y is the common ancestral trait (shared by F), and that x was derived independently at b and d – two evolutionary events. Second possibility is that x is the ancestral trait shared by F and that y was independently derived in E and B at points e and c. Another outgroup analysis is necessary to determine which hypothesis is correct.

In this case, outgroup G has the trait x. The most parsimonious description of the origin of traits x and y are that the common ancestor H had trait x (as do ancestors F and D), and that y was derived independently at e and c – two steps. If y were ancestral (meaning here that the common ancestor H had trait y) then you would have to postulate 3 steps, the independent evolution of trait x at f, b and d. But what is the best explanation for the origin of traits x and y if G had the trait y instead of x?