© 2007 Nature Publishing Group http://www.nature.com/naturemethods Microfluidics for in vivo imaging of neuronal and behavioral activity in Caenorhabditis elegans Nikos Chronis 1,2 , Manuel Zimmer 1 & Cornelia I Bargmann 1 The nematode C. elegans is an excellent model organism for studying behavior at the neuronal level. Because of the organism’s small size, it is challenging to deliver stimuli to C. elegans and monitor neuronal activity in a controlled environment. To address this problem, we developed two microfluidic chips, the ‘behavior’ chip and the ‘olfactory’ chip for imaging of neuronal and behavioral responses in C. elegans. We used the behavior chip to correlate the activity of AVA command interneurons with the worm locomotion pattern. We used the olfactory chip to record responses from ASH sensory neurons exposed to high-osmotic-strength stimulus. Observation of neuronal responses in these devices revealed previously unknown properties of AVA and ASH neurons. The use of these chips can be extended to correlate the activity of sensory neurons, interneurons and motor neurons with the worm’s behavior. How neural circuits process information to generate behavior is a fundamental question in neuroscience. To address this question, one should observe an animal in a well-controlled environment, in which a specific behavior can be generated and corresponding neuronal activity monitored. Ideally such an environment should not disturb normal neuronal function and should be able to reveal the specific neuronal circuit under study. C. elegans, with its optically accessible, stereotyped and compact nervous system, has drawn great scientific attention because of its diverse repertoire of behavioral outputs and its genetic conserva- tion with vertebrates. Initial efforts to measure activity in the C. elegans nervous system relied on electrophysiological recordings from single neurons in dissected worms 1 . The recent development of genetically encoded fluorescent calcium indicators 2 has spawned an increasing interest in optical imaging approaches that permit the tracking of calcium transients in individual neurons in vivo in intact worms 3 . Although transgenic worms that express neuron-specific indica- tors can now routinely be generated, the present methods for confining and stimulating the worm during imaging are not ideal. The typical experimental setup involves application of glue onto specific segments of the worm to achieve permanent immo- bilization on a hydrated agar pad. Fluid-filled pipettes, tempera- ture-controlled plates and sharp electrodes have been used in the past to deliver chemical, thermal and mechanical stimuli, respec- tively 4,5 . Whether the organic glue is toxic to the worm and how it influences neuronal activity are difficult to determine. Moreover, the delivery of chemical stimuli to the glued worm cannot be precisely controlled or separated from mechanical stimuli asso- ciated with fluid flow. More concerns arise when the circuit controlling locomotion is under study. The glue immobilizes the worm, not allowing muscles and stretch-receptor neurons, if any, to contract and relax normally. This mechanically restricted micro- environment might affect the function of the proprioceptive sensory neurons as well as motor neurons. Most importantly, the glue setup does not permit most behaviors to be generated, visualized, quantified or correlated to neuronal activity in real time. A system with two objectives 6 has been a welcome step toward simultaneous neuronal-behavior analysis, as has been a new system for tracking thermosensory neurons (albeit at low optical resolution) in freely moving worms 7 . Recent advances in microfabrication technology permit the construction of well-controllable microenvironments with applica- tions ranging from cell analysis to tissue engineering 8,9 . In previous studies, microfluidic delivery systems have been used to trap and stimulate single cells and embryos 10,11 . In this work we extend the applications of microfluidics to in vivo C. elegans imaging. We designed and engineered microfluidic devices for trapping and stimulating single worms while monitoring their behavior and neural function. We describe two ‘worm chips’, each one having a distinct purpose: (i) to correlate behavior and interneuron activity and (ii) to reveal stimulus-response relationships in chemosensory neurons. Using these worm chips, we show that the activity of the AVA interneurons is directly correlated to the worm’s locomotion pattern and that ASH sensory neurons have a complex multiphasic response to osmotic stimuli. RESULTS The behavior chip The first microfluidic device, the behavior chip (Fig. 1a), permitted the analysis of forward or backward worm locomotion with simultaneous recording of neuronal activity. The behavior chip consists of a worm trap, whose dimensions (1,200 mm long  70 mm wide  28 mm thick) were optimized for the size of young adult RECEIVED 1 APRIL; ACCEPTED 9 JULY; PUBLISHED ONLINE 19 AUGUST 2007; DOI:10.1038/NMETH1075 1 HowardHughes Medical Institute, The Rockefeller University, 1230 York Avenue,New York, New York 10021, USA. 2 Present address: Department of Mechanical Engineering, University of Michigan, 2350 Hayward Street, Ann Arbor, Michigan 48109, USA. Correspondence should be addressed to N.C. (chronis@umich.edu). NATURE METHODS | VOL.4 NO.9 | SEPTEMBER 2007 | 727 ARTICLES