Planarian shows decision-making behavior in response to multiple stimuli by integrative brain function

Chemosensing in addition to visual sensing and appropriate responses to various stimuli in order to find food, to escape from predators, and to communicate with other individuals, are among the most important primary functions of organisms. Planarian food-finding behavior is thought to be completely dependent on chemosensing [26], and the auricles and head margin are thought to be an important organ of chemosensing [27], although no chemosensory receptors or chemosensory neurons have been identified. Here, we established a new, tractable behavioral analysis system for chemotaxis to a food attractant, and used it to quantify the movements of planarians exposed to food as a chemoattractant. This behavioral assay system combined with RNAi may be a powerful tool for finding genes involved in chemotaxis, genes important in processing neurons in the brain, genes encoding chemosensory receptors, and other genes important in chemosensory neurons.

In this study, to evaluate the strength of a chemoattractant gradient that could be established by diffusion through a low-concentration agar gel, we made successive transfers of the chemoattractant solution, thereby successively diluting it, and directly examined whether the transferred solution was able to induce planarian chemotaxis (Figure 2). This bioassay may be useful for studying chemotaxis induced by substances for which we have no way of measuring the concentration gradient directly. When placed in this reproducible chemoattractant-gradient field, planarians spent the most time in the region containing the highest concentration of the chemoattractant (Figure 3B). Furthermore, the direction of the planarians’ movement clearly showed a tendency to move toward the region with the highest concentration of the chemoattractant (Figure 3D). These results clearly suggest that the planarian may recognize a concentration gradient of chemicals: they moved toward the higher concentration of the chemoattractant without showing increased kinesis.

In contrast, unfortunately we have not been able to prepare a texture-gradient field, and therefore we were unable to investigate how planarians recognize the difference between textured and smooth surfaces to reach a region with a preferred surface. In addition, we are unable to distinguish whether the thigmotaxis/kinesis we observed here was positive or negative (preference for a smooth surface or evasion of a textured surface), although the planarians showed a clear tendency to move to and stay for a long time in the smooth surface region (Figure 5). We speculate that planarians can attach more strongly to smooth surfaces with wider adhesion areas than to textured surfaces, and that this behavior may benefit planarians by enabling them to find safer and more stable surfaces. Consistently, in nature, planarians can usually be found under small stones or leaves, where they may avoid light, strong water flow, and predators. Moreover, we found that this feature of planarians could be utilized to arbitrarily restrict the regions where planarians moved in a field to regions without vertical walls (Additional file 1: Figure S1).

The robust regenerative ability of planarians makes them very useful for analyzing the phenotypes of amputated body fragments, such as headless animals, in order to investigate the organs regulating behaviors, since these fragments do not die. Head-amputated planarians completely lost both chemotaxis and thigmotaxis (Figures 3C, and 5A), suggesting that the head is required for both of these planarian behaviors. Furthermore, the planarian's strong regenerative ability enabled us to perform region-conditional gene knockdown by the procedure called Readyknock [11], and the resultant specific loss of neural activity in the brain by Readyknock of the synaptotagmin gene caused complete loss of both chemotaxis and thigmotaxis. These results suggest that neural activity in the brain is required for the sensing of chemical and mechanical stimuli, and/or for processing in the brain of the signals received through chemo- and thigmosensory neurons.

Note that decapitation caused a decrease of mean velocity of locomotion (Figure 3F), although the duration of the assay (600 sec) was sufficient for planarians to reach target regions from the start point in the assay chamber, and for us to judge their chemotactic behavior. Because of this decrease of mean velocity we cannot completely exclude the possibility that the decapitation-induced decrease of velocity led to the reduction of the chemotactic and thigmotactic scores. However, loss of synaptotagmin gene function in the brain did not affect their velocity of locomotion (Figure 4G), but did cause loss of these behaviors, suggesting that planarian chemotaxis and thigmotaxis are dependent on the brain in their head, and that our chemotaxis and thigmotaxis assay systems are useful for analyzing various functions of the planarian brain and of nervous-system-related genes. Previously, we reported that neural activity in the brain is required for both phototaxis and thermotaxis [7,11,28]. Our results in this study together with those previous results strongly suggest that planarian behaviors responsive to external stimuli are controlled by the brain.

Planarian higher brain function was first described in the 1950s, when Thompson and McConnell used classical conditioning experiments using light and electrical shock to show planarian's ability to learn and remember [29]. However, as this phenomenon has not been verified by other groups, McConnell's studies of planarian learning remain controversial [30,31].

The unique features of planarians’ higher brain functions have provided many insights into brain function in metazoans, and studies on planarian classical conditioning have attracted many researchers [32,33]. However, the neural pathways of planarian higher brain functions, including integration and learning, have not yet been unraveled, and even many of the original observations made decades ago have not been verified yet at the neuronal level. Moreover, it is difficult to analyze the neural mechanisms underlying higher brain function in planarians using the originally reported classical conditioning assay system utilizing light and electrical shock, because electrical stimulation might affect muscles directly. Previously, assay systems for planarian phototaxis and thermotaxis have been reported, and a number of genes and neuronal subtypes involved in sensing external stimuli and related characteristic behaviors have been identified [6,7,28]. Here we took another step forward by developing quantitative methods for chemotaxis and thigmotaxis/kinesis that enabled us to investigate the mechanisms underlying information processing and integration in the planarian brain. In order to acquire information about planarian behaviors in response to multiple simultaneous stimuli, we performed a binary competitive assay in which a planarian was given two distinct stimuli simultaneously under specific conditions in combinatory experiments. The results showed that planarians displayed an order of priority among their responses to these different stimuli (Figure 6), suggesting that the planarian brain has the ability to integrate signals coming from outside to decide appropriate planarian behaviors.

Furthermore, we showed that planarian movements vary according to the stimulus level (Figure 7E), although we did not find any light-intensity dependence of planarian phototaxis (data not shown). These results suggest that planarians do not show a simple, direct response to a stimulus, but rather integrate external stimuli and then behave suitably in response to the overall conditions.

Figure 8 shows all the planarian neural circuits predicted from this and previous studies. Several receptors for light and chemical stimuli were previously identified [23,34,35]. Regarding brain interneurons, RNAi studies of planarian glutamic acid decarboxylase (DjGAD) demonstrated the involvement of GABAergic neurons in phototaxis [28]. We found recently that planarian transient receptor potential ion channel melastatin family a (DjTRPMa) may be involved in sensing temperature, and that brain-specific DjTPH(RNAi) caused loss of thermotaxis, suggesting that serotonergic neurons in the brain are required for processing of thermosensing signals to produce thermotaxis [7]. To produce a behavior in response to a stimulus, planarian locomotion is coordinated by dopaminergic neurons [36], and motor function of body wall muscles is regulated by cholinergic neurons [37] (Figure 8A).

It seems unlikely that individual neurotransmitters are involved in only single neural circuits or behaviors [24,28,37]; rather, neurons expressing a particular neurotransmitter may have different roles according to the particular cell or position in the brain, and may accordingly construct more complex neural networks. By accumulating further data from the kinds of experiments reported here and further refining the assay systems, we should be able to unravel the neural networks or systems involved in the higher brain function in planarian. Here we found that planarian behaviors in response to stimuli could be changed flexibly according to the level of stimulation. Axons of the sensory neurons in planarian, such as the axons of the optic neurons and chemosensory neurons, project to different regions in the brain (Figure 8B,C). For example, visual neurons project to the dorso-medial region (visual center) in the brain, whereas axons of chemosensory neurons in the lateral branch neurons project to the ventro-lateral region of the brain, as visualized by immunostaining with specific antibodies (Figure 8B,C), consistent with previous results obtained using fluorescent dye tracing [2]. It has been suggested that external stimuli sensed by such organs are integrated in some region(s) of the brain, finally resulting in planarians’ behaviors. Ultimately, the information acquired by sensory neurons might be accumulated inside the brain, and then processed and integrated to transduce the signals into the activity of motor neurons (Figure 8).

The chemotaxis and thigmotaxis/kinesis assay methods established here are useful for quantitatively analyzing planarian chemotactic and thigmotactic/kinetic behaviors. Headless planarian body fragments and planarians treated by Readyknock of synaptotagmin in the brain lost both chemotactic and thigmotactic behaviors, suggesting that brain neural activity is required for the chemotactic and the thigmotactic behavior in planarian. When we tested the priority among four planarian behaviors (phototaxis, chemotaxis, thigmotaxis/kinesis, and thermotaxis) by presenting the respective stimuli to planarians simultaneously as binary stimuli, the results revealed that planarians showed a clear order of predominance of these behaviors, with chemotaxis being the strongest stimulus under our conditions. We also found that planarians showed predominance of their behavioral response to either of two simultaneously presented stimuli depending on the strengths of the two stimuli in a competitive binary stimuli assay using light and chemoattractant. Taken together, our results support the notion that external stimuli sensed by respective sensory organs are integrated in the brain and determine planarian behaviors.

We thank Dr. Elizabeth Nakajima for critical reading of the manuscript. This work was supported by a Grant-in-Aid for Young Scientists (B) to TI, The Kyoto University Research Funds for Young Scientists to TI, The Brain Science Foundation funds to TI, a Grant-in-Aid for Scientific Research on Innovative Areas to KA, a Grant-in-Aid for Creative Scientific Research to KA, and the Grants to Excellent Graduate Schools (MEXT) program of Kyoto University.

Planarian behaviors in arbitrary smooth surface region. (A) Schematic drawing of textured dish for thigmotaxis/kinesis assay. Textured regions are indicated by gray color. The other region (colored white) indicates a smooth region. (B) Heat map view of planarians’ movement. Planarians only moved on the smooth surface region, although planarians usually tend to move near the edge of a dish as a default behavior.