Information on Dan's research from his CSU East Bay website: Dr. Daniel T. Cerutti Assistant Professor Psychology Department California State University, East Bay 25800 Carlos Bee Blvd. Hayward, CA 94542 Tel. 510.885.3592 (office) Fax 510.885.2553 Email: daniel.cerutti@csueastbay.edu Several of our ongoing experiments employ the zebrafish, an important model organism of vertebrate genetics. The largest part of my research aims to understand the timing behavior of birds and fish working for delayed rewards. Although they appear to be adapted to very different niches, their timing is surprisingly similar to our own—pigeons and fish underestimate the arrival of periodic food in the same way that we underestimate the time to complete a task. My research is directed to understanding two forms of questions about this behavior, how and why? Although I was “raised” in the operant tradition of B. F. Skinner, emphasizing the study of how contingencies affect behavior, I have been inevitably drawn to the ecological question of why behavior, such as timing, takes the form it does. I believe that comparative cognitive approaches and operant methodologies can be combined to help answer some of these questions by revealing how the dynamics of operant feeding are related to niche. A basic datum in these experiments is the latency to respond for a delayed reward, wait time in technical jargon. Waiting shows a lawful relationship to delay, for example, if the delay to reward is fixed (e.g., in a fixed-interval schedule), wait is a simple proportion of delay (a relationship known as “linear waiting”). This simple finding has led me to develop two related lines of investigation. The first is concerned with the question of whether simple forms of timing can help us to understand more complex examples of behavior such as choice and conditioned reinforcement? The second is concerned with the question of the circumstances under which timing evolves; and relatedly, are there other patterns of timing that have evolved? I have been investigating these questions using traditional operant techniques in which an animal works for food by pecking a response key (pigeons) or moving a lever (fish). Some of this work has been done in collaboration with Professor John Staddon and more recently with Dr. Jeremie Jozefowiez. The pigeon is a natural choice for choice experiments because it has been a subject in these studies for over 40 years, it is a reliable performer, and a sturdy species. Fish are of interest because they are arguably the oldest vertebrates and because they show a fantastic range of specialized adaptations. The African cichlid fish provide a particularly excellent opportunity to study the evolution of behavior, such as timing, because literally thousands of different species, differing in behavior and habitat, have evolved in close proximity. I suspect that they might be the “Darwin’s finch” for understanding the ecological basis of behavior (and perhaps also a window into the evolution of timing). These research programs are described in greater detail below. Timing and Choice in Pigeons Choice. In recently published experiments on choice, we found that pigeon’s responses to different sources of reinforcement (i.e., concurrent schedules) reflect a timing process; not a response strengthening process as previously thought (Cerutti & Staddon, 2003, 2004). In summary, it now appears that what pigeons learn about reinforcers in a choice task is a simple rule of when and where. That is, when the pigeons are faced with two choices, each presenting a different rate of reinforcement, they begin responding on the choice with the shortest delay to reinforcement (i.e., the higher rate of reinforcement) since the last reinforcer, and only later respond on the choice with the longer delay to reinforcement. This kind of response timing results in a higher response rate on the alternative with the higher rate of reiforcement, but the underlying behavioral process is temporal control. Real-time computer simulations of this process reveal that the pigeon choice data are well described by a timing model that assumes a single reinforcement memory that specifies when to respond where (Cerutti, in press). The model also predicts less obvious findings such as a power relation between relative response rates and relative reinforcement rates, undermatching (a tendency toward indifference between alternatives), and differential effects of different reinforcement distributions (e.g., less undermatching with geometric versus arithmetic inter-reinforcement intervals, for example). Subsequent empirical studies with pigeons confirm these predictions and further suggest that behavior on concurrent schedules can be understood better as a temporal-stochastic process in which (a) the overall reinforcement delay determines the overall rate of behavior and (b) the individual schedules’ reinforcement delays determine the allocation of responses to each alternative (Jozefowiez, Cerutti, & Staddon, in press). Ongoing experiments are attempting to test this view and provide data from which more rigorous models can be derived. We are now exploring how our model can accommodate a wide range of findings, from those typically understood to be conditioned-reinforcement effects to self-control phenomena (review in Staddon & Cerutti, 2000). Experiments on conditioned reinforcement. Much behavior is not maintained by immediate reward, but by the promise of a later reward—signaled-delayed reinforcement. The usual interpretation of signals that maintain behavior in this way is that they are Pavlovian conditioned stimuli. Although there are some examples where this is clearly the case, there are others that defy a Pavlovian account. For example, in a two-link chain schedule, responding in the presence of stimulus S1 (e.g., a red light) produces stimulus S2 (e.g., a green light), and responding in the presence of S2 produces primary reinforcement (e.g., food). Given this arrangement, it is conventionally supposed that responding in S1 is reinforced by S2 because S2 is a conditioned reinforcer. But if, as I believe, temporal control of any sort is common to all schedules of reinforcement, it should also play a role in this situation. Preliminary experiments show that it does. For example, if S1 and S2 represent fixed-interval schedules, pausing in S1 is a function of the delay to primary reinforcement (i.e., the duration of S1 + S2), not the delay to the onset S2 (as implied by Pavlovian theory). Findings like these have led me to explore the evidence for timing in procedures that arrange choices between signaled-delayed reinforcers (technically, concurrent-chain schedules). Preliminary findings show that timing is not always sufficient to explain all concurrent-chain findings. In ongoing tests of choices between conditioned reinforcers, I have found that pigeons sometimes choose on the basis of time to the onset of the conditioned reinforcer (S2) and sometimes choose on the basis of the time to reinforcement (S1+S2). This switch between choice strategies is a novel and potentially important finding and I am now directing my attention to understanding the causes. A number of experiments to test ideas about this process have been submitted in a grant to NIMH. Timing and Niche in Fish Timing in a predator. A tacit assumption of research on timing is that time is an ecologically important dimension of natural contingencies. If so, animals living under vastly different ecological constraints should show some degree of specialization in timing. I have spent a good chunk of the last few years refining apparatus and procedures to study timing in fish in order to test ideas about temporal adaptations. The present apparatus consists of an internally illuminated lever, suspended directly beneath a pellet feeder that dispenses dried fish pellets. Preliminary experiments have been done with the Oscar (Astronotus ocellatus), a wait-and-ambush piscivore from the Amazon River. The Oscar was chosen because it is readily available and has a reputation among hobbyists for “intelligence.” It also provides an enormous contrast with rats and pigeons, whose timing behavior is most studied, which are omnivorous scatter foragers. Oscars show the familiar pause function in fixed-interval schedules, pausing a proportion of the fixed-interval schedule duration, although the slope is shallow and the intercept in nonzero (Cerutti, Talton, & Staddon, in preparation). Dynamical tests, however, showed a pattern opposite to that seen in rats and pigeons, with more rapid adjustments to increasing than decreasing delays to food. This made sense given the Oscar’s wait-and-ambush feeding strategy, and the next experiment sought to provide a better look at its divergence in timing style. An obvious test of the Oscar’s specialized timing ability seemed to be the differential-reinforcement-of-low-rate (DRL) schedule in which reinforcement is arranged for a response after a minimum time has elapsed since the last response. A significant finding with the Oscar was that it is able to learn to withhold a responses on the DRL schedule for very long time intervals (I tested times up to 120 s in the Oscar). In contrast, rats and pigeons can only withhold responses for about 20 seconds. Beyond that, timing breaks down. A plausible interpretation of the Oscar finding is that waiting (i.e. pausing) in such a predator is a dimension of behavior that is readily reinforcible. The case for pigeons and rats seems to be different, perhaps because waiting is not a common feature of their natural foraging behavior. Predators versus grazers. The findings obtained with Oscars are exciting, but do not constitute strong proof of that the timing “faculty” for any given species is tuned to ecological constraints. There is after all the possibility that other fish also show these wait dynamics. This question forms the basis for current experiments that study timing in closely related species of African cichlids, some of which are wait-and-ambush predators (Nibochromis Livingtonii) and others which are bottom grazers (Pseudotropheus zebra). The choice of these species is carefully considered: I anticipated that the wait-and-ambush species will behave like Oscars, showing the ability to learn to withhold responses, while the grazers will behave like rats and pigeons, showing the inability to learn to withhold responses for more than a few seconds. Thus far the findings have confirmed my predictions. I believe that these experiments, and additional related studies I have planned, will provide a window into the evolution of timing. I also anticipate that the use of closely related species will permit a correlational analysis of brain morphology related to timing, for which I hope to find collaborators. This research I find particularly exciting and will seek funding from NSF and NIMH as my preliminary studies are published. Zebrafish cognition. Zebrafish have become the most important model for the molecular processes of neural development. The reason is that they have a clear chorion, permitting visual monitoring of their nervous system during development—its genome has been sequenced, knockout zebrafish are available, and research has examined the effects gene expression by a variety of molecular operations. Despite all of this biological research, little is know about typical zebrafish behavior or how genetic manipulations affect behavioral processes. The problem of studying zebrafish behavior with operant methodology is daunting—zebrafish are only about 3 cm in length and weigh but a few grams. For these reasons most research has been done with mazes, but I have worked hard this last year to produce a subminiature lever apparatus for zebrafish much like that used with rats. The advantage of such an apparatus is that it allows us to more precisely determine the nature of the animal’s behavior by manipulating stimulus-response contingencies. For example, I just recently produced what appears to be autoshaped responding (Pavlovian conditioning) in four zebrafish and I am now conducting experiments to determine if the behavior is true or pseudo conditioning. Future experiments with zebrafish will be directed to studying their instrumental behavior, discrimination, timing, and memory. I believe that this work should be of great interest to zebrafish researchers concerned with genetic expression and I am currently seeking funding from NIH (with Dr. Edward Levin, Department of Psychiatry and Behavioral Sciences, and Dr. Elwood Linney, Department of Molecular Biology) for additional behavioral work as my preliminary studies are published.