about the context in which behaviours occur. These neurons also receive dopamine-medi-ated inputs from the midbrain that report on the significance of the behaviour — for exam-ple, whether it yields an unexpected reward.

When something important happens, the striatum stimulates circuits that project via the thalamus to the prefrontal cortex; this leads to the formation of memories that can later guide the planning and control of behav-iour. Abnormalities in synaptic function within frontal–striatal–thalamic loops could lead to unintended behaviours or even unintended thoughts. The main significance of the Sapap3-deficient mouse is that it represents a model of how such a pathological process might work. In this case, abnormal synaptic function causes grooming behaviours that are repetitive enough to produce skin injuries and unrespon-sive enough to control by the prefrontal cor-tex that they continue even after the injuries have occurred.

The mouse model described by Welsh and colleagues will excite those interested in the pathophysiology of OCD or in developing treatments for it. But we must retain a healthy caution. Genetic animal models of psychiatric disorders have been produced by inbreeding for desired traits9, by introducing human disease genes10, by mutating existing animal genes11 or, as in this case, by observing an unexpected behavioural response in a genetically engi-neered animal. It is highly unlikely that such animal models will ever recapitulate human psychiatric disorders in their entirety. However, as in this case, it is possible to model significant aspects of these diseases.

The core symptom of OCD — unwanted intrusive thoughts — cannot be mimicked in mice, at least not in any obvious way. Instead, Sapap3-deficient mice show excessive groom-ing that continues past the point of self-harm. This is reminiscent, perhaps, of compulsive hand-washing, which in patients with OCD represents a never fully successful attempt to neutralize fears of contamination.

In OCD patients, the main cause of anxiety is the unwanted intrusive thoughts. The suffer-ers are anxious because they cannot be certain that the door is locked, the gas has been turned off, or that they are free of dreaded microbes. The anxiety-like behaviours observed in these mice may also resemble OCD, but this requires a stretch of the imagination.

Even if we can gain assurance with addi-tional research that the behaviours observed in Sapap3-deficient mice reflect abnormalities in circuits that produce human symptoms, we cannot assume that OCD-related conditions in humans involve variations in this gene. These disorders, like other major psychiatric dis-eases, seem to be heterogeneous with complex underpinnings — probably involving several genes — that, in interaction with developmental and environmental factors, could lead to abnor-malities in frontal–striatal–thalamic circuits.

Despite these reservations, the work of Welsh

et al.2 sharpens our focus on frontal–striatal–thalamic circuits both in human patients and in animal models of OCD. It also gives us a compelling clue that the compulsive behaviour associated with this condition is due to a syn-aptic abnormality in these neural loops. Such cellular insight should aid an understanding of OCD at a molecular level. ■

Steven E. Hyman is in the Department of Neurobiology, Harvard Medical School, Massachusetts Hall, Cambridge, Massachusetts 02138, USA.e-mail: seh@harvard.edu

QUANTUM PHYSICS

Wave goodbyeLuis A. Orozco

When measuring photons, it’s a case of ‘wanted, dead’ — catching them alive is not an option. But we can observe how a superposition of many photon waves progressively collapses as it interacts with a beam of atoms.

Earlier this year, a team from the Ecole Nor-male Supérieure in Paris recorded jumps of light heralding the birth and death of a pho-ton trapped in a cavity1. As they describe in this issue (Guerlin et al., page 889)2, the same researchers have now performed a similar, more complex trick — recording exactly how a coherent state of many photons collapses as it is measured.

A measurement process differs fundamentally between the classical and quantum worlds. In the classical realm, there is no explicit limitation on a measurement’s accuracy. In the quantum domain, by contrast, accuracy is constrained by the Heisenberg uncertainty principle: a measurement will produce a definite result, but one whose value is distributed according to the laws of probability. What is more, the meas-ured object will itself be fundamentally altered by the measurement. Thus, the clicking sound produced when a photon is caught by a detector says two things: yes, a particle was detected; but sorry, the way you detected it killed it, and its energy was converted into an electric pulse.

But the quantum world has more subtle states to investigate than a single photon. Pho-tons, or the probabilistic wavefunctions associ-ated with them, can add together, or superpose. If they superpose coherently (in phase), their combined wavefunction begins to look like a classical wave. This coherent electromagnetic field is the complex beast whose collapse was monitored by Guerlin et al.2.

But how did they achieve this feat, given the difficulties of measuring a quantum object without instantly destroying it? The authors’ ‘quantum non-demolition measurements’ in a cavity quantum-electrodynamical (QED) system required profound understanding of quantum mechanics, continuous theoretical

elucidation of subtle details of cavity QED, and unprecedented dedication in realizing a simple theoretical model in the laboratory. This model3 first required the development of a pair of superconducting mirrors for the walls of the cavity whose losses are low enough that light remains captured between them for the length of time it would take the light to circle Earth at the Equator.

The second pivotal ingredient is individual rubidium atoms in a ‘Rydberg’ state in which one electron is highly excited. These atoms are like little planetary systems, with the excited electrons on a distant orbit around a remote atomic nucleus. They can oscillate between two different excited states, and the regularity of this oscillation makes them excellent time-keepers. The frequency of that oscillation is easily disturbed in the presence of light — to the extent that it can be used to detect the pres-ence of a single photon non-destructively1.

And so Guerlin et al.2 prepared a coherent state of a microwave electromagnetic field in their cavity that contained up to seven photons, and then sent in the Rydberg atoms. They made sure that these atomic clocks had their ‘hands’ — their initial phase — set to one of eight pos-sible values. Each of these settings is a possible answer to the question “How many photons are there in the cavity?” The interaction with the electromagnetic field modifies the setting of the hands according to the answer, from none to seven. The atoms leave the cavity, and the new direction of the hands is recorded.

Guerlin and colleagues’ measurement method had to fulfil certain conditions for success4. First, the interaction of the quantum probe (the atom) with the remaining, classi-cal part of the measuring device (the ‘click’ of the detector) must not begin until the

1. Chamberlain, S. R., Menzies, L., Sahakian, B. J. & Fineberg, N. A. Am. J. Psychiatry 164, 568–574 (2007).

2. Welch, J. M. et al. Nature 448, 894–900 (2007).3. Hollander, E., Kim, S., Khanna, S. & Pallanti, S. CNS Spectr.

12 (suppl. 3), 5–13 (2007). 4. Merikangas, K. R. & Risch, N. Am. J. Psychiatry 160,

625–635 (2003). 5. Robinson, D. et al. Arch. Gen. Psychiatry 52, 393–398

(1995).6. Pujol, J. et al. Arch. Gen. Psychiatry 61, 720–730 (2004).7. Bloch, M. H., Leckman, J. F., Zhu, H. & Peterson, B. S.

Neurology 65, 1253–1258 (2005). 8. Remijnse, P. L. et al. Arch. Gen. Psychiatry 63, 1225–1236

(2006). 9. Bice, P. J. et al. Behav. Genet. 36, 248–260 (2006).10. Shahbazian, M. et al. Neuron 35, 243–254 (2002).11. Clapcote, S. J. et al. Neuron 54, 387–402 (2007).

872

NEWS & VIEWS NATURE|Vol 448|23 August 2007

������������������� � � ���� �������������

50 &

100

YEA

RS A

GOinteraction of the atom with the electromag-

netic field is complete. Second, the interaction with the classical part must not contribute significantly to the total error of the measure-ment. Finally, there should be no fundamental constraint on the precision of a measurement: in quantum-mechanical terms, the phase of the electromagnetic field (which according to the uncertainty principle is the ‘canonical conjugate’ of the number of photons) should be maximally perturbed by the measurement in order to measure the photon number accurately.

Each individual measurement in this pro-cess gives only one answer about the number of photons in the coherent state. As the meas-urement process is started, these answers are uniformly (randomly) distributed among the possible values, but as the measurement progresses and more and more atoms are sent through the light field, the answer con-verges (collapses) to a particular value. Start-ing the process all over again with a freshly trapped, but identically prepared, light field, the measurement converges again to a different number. By repeating the meas-urement process again and again, the distri-bution of the number of photons in the initial

coherent state can be reconstructed (Fig. 1).The ability to follow a quantum measure-

ment stage by stage as the wavefunction col-lapses opens up broad avenues of approach to fundamental questions that were previously accessible only in simulations. The process of measurement might now be manipulated to steer a system to a particular final state using quantum feedback or to explore conditional quantum dynamics — using the outcome of a measurement to determine a system’s evolution. The Parisian team is uniquely placed to advance farther into the quantum jungle, a largely unex-plored area where new species of Schrödinger’s cat rub shoulders with exotic ‘quantum butter-flies’ — flighty states that flutter between values when disturbed by measurement, only to settle down to rest again. ■

Luis A. Orozco is at the Joint Quantum Institute, Department of Physics, University of Maryland, College Park, Maryland 20742, USA.e-mail: lorozco@umd.edu

1. Gleyzes, S. et al. Nature 446, 297–300 (2007).2. Guerlin, C. et al. Nature 448, 889–893 (2007).3. Raimond, J. M., Brune, M. & Haroche, S. Rev. Mod. Phys.

73, 565–582 (2001). 4. Braginsky, V. B. & Khalili, F. Y. Quantum Measurement

(Cambridge Univ. Press, 1995).

Figure 1 | State surveillance. a, As a stream of rubidium atoms passes through a mirror-walled cavity filled by a coherent light field in Guerlin and colleagues’ experiment2, the atomic ‘clocks’ are disturbed according to the number of photons present in the light field. The probability for different photon numbers is initially evenly distributed between the possible outcomes (none to seven photons), but ‘collapses’ to a definite value as more and more atoms pass through — to five, in this case. b, A different measurement of the same state collapses to a different photon number — seven. c, By repeating the same measurement over and over again, an overall (poissonian) probability distribution for the number of photons in the initial coherent light field can be gathered. (Figure modified from ref. 2.)

654321 70No. of photons

0.05

0.10

0.15

c

a b

Prob

abili

ty

0.00

1.0

0.5

Prob

abili

ty

1.0

0.5

Prob

abili

ty

020

4060

80100

No. of atoms

0

No. of photons1

23

4567

020

4060

80100

No. of atoms

0

No. of photons1

23

4567

50 YEARS AGOOn the Beach by N. Shute — The theme of “On the Beach” is the extinction of the human race resulting from an atomic war. Everybody dies. Just that. When the novel opens in Melbourne, 1962, nobody is alive in the northern hemisphere. Movements of the atmosphere are steadily carrying lethal particles southwards. Mr Shute has deployed his remarkable imagination as engineer, naval officer and storyteller… On his immense popular following this book can only inflict a haunting distress: one takes off one’s hat to him. Mr Shute has limited his canvas: he has taken only five main characters—five very ordinary people at that; and he has pitched the emotional and dramatic tone invariably low. Until the moment when the radiation sickness comes on, everybody sticks to the tamest of domestic preoccupations; and then quietly takes a suicide pill. The effect is hypnotic and also odd. Mr Shute’s world ends not, as the epigraph warns, with a whimper, but with a stoical silence, movingly impressive…The moral in so far as Mr Shute states it explicitly, comes via his characters thus: “Maybe we’ve been too silly to deserve a world like this”… From Nature, 24 August 1957.

100 YEARS AGOAccording to Engineering, an Australian record in wireless telegraphy has been achieved by the successful transmission of messages from H.M.S. Challenger, one of the Australian squadron at present stationed in Hobson’s Bay, to the flagship Powerful, which at the time was moored in Farm Cove, Port Jackson. The Challenger was in communication with the flagship by means of wireless telegraphy the whole of her voyage. The longest message was one flashed over a distance of 410 miles in a direct line, and this constitutes an Australian record, as previously never more than 240 miles had been achieved by warships on the Australian station.From Nature, 22 August 1907.

873

NEWS & VIEWSNATURE|Vol 448|23 August 2007

������������������� � � ���� �������������