rhodopsin-like photopigment in adult eyes preserved by rapid freezing suggests that the rhabdoms of B. thermydron beyond the first juvenile crab stage may be resistant to the chemical fixation used for anatomical analysis. Rapid freezing of crabs collected remotely under low light conditions at infrequently visited vents may be required to determine the in situ state of adult and late- juvenile eyes. Because related shallow-water brachyuran crabs do not undergo visual metamorphosis12,21, the metamorphosis of the eyes of B. thermydron described here appears to be a specific adaptation for life at the vents. A Methods Animals Specimens were collected live under artificial illumination from DSV Alvin. Megalopae, collected at the vents, were identified by morphological and genetic analysis as described elsewhere13. Zoea larvae were hatched in darkness on board ship from a freshly collected ovigerous B. thermydron. Larvae were transported back to land in darkness, sorted briefly under ambient lighting, and returned to darkness before use. Microscopy Upon Alvin���s return to the surface (under natural light), whole juvenile crabs and excised adult eyes were submersed in fixative (5% paraformaldehyde w/v, 0.8% glutaraldehyde v/v, 4.5% sucrose w/v, 3% NaCl w/v in 0.1 M phosphate buffer pH 7.2) under ambient illumination aboard RV Atlantis. Larvae were fixed by immersion under ambient lighting in Lancaster, Pennsylvania, 17 d after hatching. Adults and juveniles were stored in fixative at 4 8C in darkness until they were processed for microscopy on land by using standard methods10. Sample sizes: zoea (n �� 7 eyes from 7 animals) juvenile crab stage 1 (n �� 6 eyes from 3 animals) juvenile crab stage 3 (n �� 4 eyes from 2 animals) adult (n �� 4 eyes from 2 animals). Microspectrophotometry Animals were kept in darkness for several days before being used. Adult eyes were dissected out, quick-frozen using cryogenic spray, and sectioned at 14mm. Larvae and juveniles were frozen whole, but otherwise treated like the adults. Sections were collected on coverslips, mounted in marine crustacean Ringer solution containing 2.5% glutaraldehyde v/v (to enhance photobleaching), and placed in the microspectrophotometer. Photoreceptors were selected for scanning in dim, red light, and scanned using a beam (1.5���5mm) placed in each rhabdom. Rhabdoms were scanned twice: first when fully dark-adapted, and subsequently after being photobleached for several minutes with bright, white light. The difference between the scans was taken to be the spectrum of the visual pigment. Between 3 and 32 individual rhabdoms (depending on the developmental stage under study and the quality of the material) were scanned, bleached, and averaged for spectral characterization. Averaged scans were fitted mathematically with Stavenga templates22, using a least-squares procedure, to determine their characteristic wavelengths of maximum absorption. Received 24 July accepted 10 September 2002 doi:10.1038/nature01144. 1. Grassle, J. F. Hydrothermal vent animals: distribution and biology. Science 229, 713���717 (1985). 2. Rona, P. A., Klinkhammer, G., Nelsen, T. A., Trefry, J. H. & Elderfield, H. Black smokers, massive sulphides and vent biota at the Mid-Atlantic Ridge. Nature 321, 33���37 (1986). 3. Van Dover, C. L. Hydrothermal Vents and Processes (eds Parson, L. M., Walker, C. L. & Dixon, D. R.) Special Publication No. 87 257���294 (Geological Society, London, 1995). 4. Van Dover, C. L. et al. Biogeography and ecological setting of Indian Ocean hydrothermal vents. Science 294, 818���823 (2001). 5. Herring, P. J. & Dixon, D. R. Extensive deep-sea dispersal of postlarval shrimp from a hydrothermal vent. Deep-Sea Res. I 45, 2105���2118 (1998). 6. Pond, D., Dixon, D. & Sargent, J. Wax-ester reserves facilitate dispersal of hydrothermal vent shrimps. Mar. Ecol. Prog. Ser. 146, 289���290 (1997). 7. Van Dover, C. L., Reynolds, G. T., Chave, A. D. & Tyson, J. A. Light at deep-sea hydrothermal vents. Geophys. Res. Lett. 23, 2049���2052 (1996). 8. Renninger, G. H. et al. Sulfide as a chemical stimulus for deep-sea hydrothermal vent shrimp. Biol. Bull. 189, 69���76 (1995). 9. Van Dover, C. L., Szuts, E. Z., Chamberlain, S. C. & Cann, J. R. A novel eye in ���eyeless��� shrimp from hydrothermal vents of the Mid-Atlantic Ridge. Nature 337, 458���460 (1989). 10. O���Neill, P. J. et al. The morphology of the dorsal eye of the hydrothermal vent shrimp, Rimicaris exoculata. Vis. Neurosci. 12, 861���875 (1995). 11. Jinks, R. N. et al. Sensory adaptations in hydrothermal vent shrimps from the Mid-Atlantic Ridge. Cah. Biol. Mar. 39, 309���312 (1998). 12. Cronin, T. W. & Jinks, R. N. Ontogeny of vision in marine crustaceans. Am. Zool. 41, 1098���1107 (2001). 13. Epifanio, C. E., Perovich, G., Dittel, A. I. & Cary, S. C. Development and behavior of megalopa larvae and juveniles of the hydrothermal vent crab Bythograea thermydron. Mar. Ecol. Prog. Ser. 185, 147���154 (1999). 14. Land, M. F. The sight of deep wet heat. Nature 337, 404 (1989). 15. Gaten, E., Herring, P. J., Shelton, P. M. J. & Johnson, M. L. Comparative morphology of the eyes of postlarval bresiliid shrimps from the region of hydrothermal vents. Biol. Bull. 194, 267���280 (1998). 16. Herring, P. J., Gaten, E. & Shelton, P. M. J. Are vent shrimps blinded by science? Nature 398, 116 (1999). 17. Frank, T. M. & Case, J. F. Visual spectral sensitivities of bioluminescent deep-sea crustaceans. Biol. Bull. 175, 261���273 (1988). 18. White, S. N., Chave, A. D. & Reynolds, G. T. Investigations of ambient light emission at deep-sea hydrothermal vents. J. Geophys. Res. 107, EPM 1���13 (2002). 19. Tapley, D. W., Buettner, G. R. & Shick, J. M. Free radicals and chemiluminescence as products of the spontaneous oxidation of sulfide in seawater, and their biological implications. Biol. Bull. 196, 52���56 (1999). 20. Bennett, J. T. & Turekian, K. K. Radiometric ages of brachyuran crabs from the Galapagos spreading- center hydrothermal ventfield. Limnol. Oceanogr. 29, 1088���1091 (1984). 21. Cronin, T. W., Marshall, N. J., Caldwell, R. L. & Pales, D. Compound eyes and ocular pigments of crustacean larvae (Stomatopoda and Decapoda, Brachyura). Mar. Freshwat. Behav. Physiol. 26, 219���231 (1995). 22. Stavenga, D. G., Smits, R. P. & Hoenders, B. J. Simple exponential functions describing the absorbance bands of visual pigment spectra. Vision Res. 33, 1011���1017 (1993). 23. Williams, A. B. A new crab family from the vicinity of submarine thermal vents on the Galapagos Rift (Crustacea: Decapoda: Brachyura). Proc. Biol. Soc. Wash. 93, 443���472 (1980). Acknowledgements We thank the captain and crew of the RVAtlantis, the DSVAlvin group, and members of the Epifanio laboratory for animal collection. We thank J. J. McDermott for comments on the manuscript. This work was supported by the National Science Foundation (A.I.D., C.E.E. and T.W.C.) and by Franklin and Marshall College. Competing interests statement The authors declare that they have no competing financial interests. Correspondence and requests for materials should be addressed to R.N.J. (e-mail: rjinks@fandm.edu). .............................................................. Neurons in medial prefrontal cortex signal memory for fear extinction Mohammed R. Milad & Gregory J. Quirk Department of Physiology, Ponce School of Medicine, Ponce, Puerto Rico 00732 ............................................................................................................................................................................. Conditioned fear responses to a tone previously paired with a shock diminish if the tone is repeatedly presented without the shock, a process known as extinction. Since Pavlov1 it has been hypothesized that extinction does not erase conditioning, but forms a new memory. Destruction of the ventral medial pre- frontal cortex, which consists of infralimbic and prelimbic cortices, blocks recall of fear extinction2,3, indicating that medial prefrontal cortex might store long-term extinction memory. Figure 2 Absorbance spectra for B. thermydron visual pigments. Coloured curves are Stavenga-type rhodopsin templates22. Top, the zoeal visual pigment is fitted best by a template with l max �� 447 nm (n �� 6 scans). Middle, the megalopa, the last instar before the juvenile crab, possesses a visual pigment with l max �� 479 nm (n �� 3 scans). Bottom, adult eyes contain a red-shifted rhodopsin-like pigment with l max �� 489 nm (n �� 18 scans). letters to nature NATURE | VOL 420 | 7 NOVEMBER 2002 | www.nature.com/nature 70 �� 2002 Nature Publishing Group

Here we show that infralimbic neurons recorded during fear conditioning and extinction fire to the tone only when rats are recalling extinction on the following day. Rats that froze the least showed the greatest increase in infralimbic tone responses. We also show that conditioned tones paired with brief electrical stimulation of infralimbic cortex elicit low freezing in rats that had not been extinguished. Thus, stimulation resembling extinc- tion-induced infralimbic tone responses is able to simulate extinction memory. We suggest that consolidation of extinction learning potentiates infralimbic activity, which inhibits fear during subsequent encounters with fear stimuli. Rats were given auditory fear conditioning in a 2-day experi- ment2,4. The conditioned stimulus was a 30-s tone, and the uncon- ditioned stimulus was a 0.5-s foot-shock that terminated simultaneously with the tone. On day 1, rats received five tones (habituation phase) followed immediately by five tones paired with foot-shock (conditioning phase). After 1 h, rats received 20 tones without foot-shock (extinction phase). On day 2, an additional 15 tones were given to test for recall of extinction learning. Freezing to the tone increased to 75% during the conditioning phase, and decreased to 10% by the end of the extinction phase (Fig. 1b). Twenty-four hours later, rats showed 35% freezing, which reflects good recall of extinction learning compared with the 80% freezing in animals that were not extinguished (data shown in Fig. 4b). A total of 74 neurons were recorded from the medial prefrontal cortex (mPFC) across the 2 days. Cells were located in the infra- limbic cortex (IL, n �� 31, see Fig. 1a), prelimbic cortex (PL, n �� 25) or medial orbital cortex (MO, n �� 18). The initial spontaneous firing rate of all cells was 4.6, 5.5 and 4.2 Hz for IL, PL and MO, respectively. Spontaneous rates did not change significantly across the different phases of the experiment (P . 0.3, one-way analyses of variance (ANOVAs) with repeated measures). We therefore focused on the tone-elicited activity of neurons. Figure 1c shows an example of a tone-responsive neuron in IL. No tone-elicited activity was observed during habituation, and cells remained unresponsive to tones during the conditioning phase. This agrees with previous studies showing that lesions of mPFC do not prevent acquisition of fear conditioning2,3,5. Thus, unlike more dorsal parts of mPFC6, IL does not signal the acquisition phase of fear conditioning. IL cells continued to be unresponsive to tones during extinction training on day 1. By day 2, however, robust tone-elicited activity in IL was visible from the start of the extinction phase. The increase in the neuronal tone response from day 1 to day 2 was inversely correlated with the spontaneous recovery of freezing on day 2 (r �� 20.73, P , 0.01 Fig. 2a). Rats with the largest increase in IL tone responses showed the least freezing. To investigate further the relationship between IL activity and freezing, we divided the rats into two groups: those showing less than 50% spontaneous recovery of freezing and those showing more than 50% spontaneous recovery. As shown in Fig. 2c, IL activity was significantly increased 100���400 ms after tone onset in low-recovery rats but not in high- recovery rats. IL tone responses on day 2 are unlikely to reflect recall of conditioning because they were lowest in animals that showed the highest freezing. A more parsimonious explanation is that IL tone responses reflect extinction memory. Extinction-induced tone responses were not observed in nearby PL or MO (see Fig. 2c), indicating a high degree of anatomical specificity in the ability of mPFC to signal extinction memory. Figure 1 Tone response of a representative neuron in medial prefrontal cortex (infralimbic area, IL) during acquisition and extinction of conditioned fear. a, Unit-recording electrode in IL. b, Freezing to the tone shown in blocks of two trials for 24 rats. Freezing was low on day 2, indicating good recall of extinction learning. c, Waveforms and post- stimulus time histograms (PSTHs) showing the tone-elicited activity of a representative IL neuron in each phase of the experiment (10 trials each, bin �� 100 ms). Dashed line indicates tone onset. IL neurons only signalled the tone 24 h after extinction training. Cond, conditioning habit., habituation. Figure 2 IL tone responses are correlated with spontaneous recovery of freezing after extinction. a, Scatter plot showing the change in IL tone response across days versus the percentage recovery of freezing on day 2. Firing rate 0���400 ms after tone onset was compared to pre-tone baseline rate with z-score. Each point represents the averaged response of all recorded neurons in each rat. Filled squares, low-recovery group (,50% 7 rats, 19 cells) open squares, high-recovery group (.50% 5 rats, 12 cells). r �� 20.73 P , 0.01. b, Recording sites in IL, PL and MO. c, Group PSTHs showing tone responses of neurons from high-recovery (IL,12 cells PL,12 cells MO, 9 cells) and low-recovery (IL,19 cells PL,13 cells MO, 9 cells) groups on day 2. The bin size was 50 ms. letters to nature NATURE | VOL 420 | 7 NOVEMBER 2002 | www.nature.com/nature 71 �� 2002 Nature Publishing Group