The strategies by which foraging predators decide when to redirect their gaze influence both prey detection rates and the prey's ability to detect and avoid predators. We applied statistical analyses that have been used to study neural decision-making for gaze redirection in primates to 3 species of predatory birds with different sizes, visual systems, habitats, and hunting behaviors: the Northern Goshawk (Accipiter gentilis), Cooper's Hawk (A. cooperii), and Red-tailed Hawk (Buteo jamaicensis). The timing of head saccades was measured during visual searches using field video recordings of foraging raptors, and during a variety of behaviors using a miniature camera mounted on the head of a Northern Goshawk. The resulting statistical distribution of latencies (time between successive head saccades) was compared to predictions from various models proposed to describe visual search strategies. Our results did not support models that assume a constant probability of gaze redirection per unit time, a constant time for “giving up” on the visual search, or an initial setup time before visual search initiation. Instead, our data were fit best by a log-normal distribution, consistent with the raptors stochastically changing their gaze direction on the basis of accumulated environmental information. Specifically, this suggests that saccade initiation arises from a neural computation based on detection of a threshold level of a dynamically updated decision signal that encodes noisy sensory data, similar to the processes inferred from previous studies of visual search strategies in primates. The only significant between-species difference we found was a slower mean gaze-redirection rate for 2 larger species compared to the Cooper's Hawk, even though the latter has hunting behavior and maneuverability similar to that of the Northern Goshawk. Head-saccade latencies measured for a Northern Goshawk during different behaviors showed that the bird changed gaze direction significantly less frequently, on average, while perched than while in motion.
Predatory birds must scan a complex three-dimensional environment to locate, track, and pursue prey. To do so, they change the direction of their gaze via saccades (rapid motions of the head or eyes) that alternate with long periods of visual fixation (Wallman and Letelier 1993, Land 2015). Because saccadic head motions are discrete, unambiguous events resolvable by field video recordings with good temporal resolution compared to the relevant behavioral timescales, they also enable the study of visual search dynamics during foraging and hunting (Land 1999b, Gall and Fernández-Juricic 2010, O'Rourke et al. 2010b). Past studies of visual searches by avian predators have explored the influence of factors such as hunting behavior (O'Rourke et al. 2010b), prey crypticity (Beauchamp and Ruxton 2012), the formation of an effective search image (Anderson et al. 1997, Alpern et al. 2011, Hein and McKinley 2013), and perch height (Andersson et al. 2009, Tomé et al. 2011). Specific models proposed for foraging searches by predatory birds have suggested that they redirect their attention at an average (possibly optimal) rate or do so with an average probability per unit time (Fitzpatrick 1981, Stephens and Krebs 1986) to optimize search efficiency and, hence, net energy intake. By contrast, it has been hypothesized that the mechanisms governing saccades in primates evolved early in evolutionary history, driven by the needs of both predator and prey for effective searching and unpredictable gaze shifts, to avoid signaling their next moves (Carpenter 1999). Indeed, members of many taxa, including birds, have been shown to react to the gaze direction of potential predators (Davidson et al. 2014), which suggests that unpredictable saccade timing can confer a fitness benefit. Each of these hypothesized models makes specific predictions for the distribution of saccade latencies (intersaccade intervals). For example, if saccades are initiated with a constant probability per unit time, their latency distribution should be a decaying exponential, while one would expect a normal distribution if there is a preferred (possibly optimal) saccadic latency. Therefore, we decided to measure this distribution for naturally foraging predatory birds as a proxy for determining the different mechanisms of neural decision-making that could be at work in this system.
Quantitative neurophysiological models of decision-making have been used to explain the timing of eye saccades in humans and other primates (Carpenter 2012), with support from empirical data at the neural level (Gold and Shadlen 2001). These models have been used to explain several surprising features of primate eye saccades, such as the broad right-skewed probability distribution of saccade latencies and the fact that mean latencies are appreciably longer than the minimum time required for sensing visual stimuli and muscle activation (Carpenter 1999). The decision to initiate a saccade is assumed to be based on detection of a threshold level, ST, of a decision signal, S, based on sensory data. This decision signal is assumed to start at an initial level, So, based on prior knowledge and to rise linearly at a rate, R, based on the output of a neural network that processes incoming sensory inputs. Such “rise-to-threshold” models provide a mechanism that introduces randomness in the decision-making process itself, in addition to randomness generated by noise arising from the environmental and sensory system (Carpenter and Williams 1995). The decision signal, S, is assumed to encode in some way the probability of a hypothesis being true (e.g., for a spontaneous saccade, “There a