of neural activity
Edward S Boyden1, Feng Zhang1, Ernst Bamberg2,3, Georg Nagel2,5& Karl Deisseroth1,4
Temporally precise, noninvasive control of activity in well-
defined neuronal populations is a long-sought goal of systems
neuroscience. We adapted for this purpose the naturally
occurring algal protein Channelrhodopsin-2, a rapidly gated
light-sensitive cation channel, by using lentiviral gene delivery in
combination with high-speed optical switching to photostimulate
mammalian neurons. We demonstrate reliable, millisecond-
timescale control of neuronal spiking, as well as control of
excitatory and inhibitory synaptic transmission. This technology
allows the use of light to alter neural processing at the level of
single spikes and synaptic events, yielding a widely applicable
tool for neuroscientists and biomedical engineers.
Neural computation depends on the temporally diverse spiking pat-
and demonstrate heterogeneous wiring properties within neural net-
works. Although direct electrical stimulation and recording of neurons
in intact brain tissue have provided many insights into the function of
circuit subfields (for example, see refs. 1–3), neurons belonging to a
specific class are often sparsely embedded within tissue, posing funda-
mental challenges for resolving the role of particular neuron types in
information processing. A high–temporal resolution, noninvasive,
genetically based method to control neural activity would enable
elucidation of the temporal activity patterns in specific neurons that
drive circuit dynamics, plasticity and behavior.
Despite substantial progress made in the analysis of neural network
geometry by means of non–cell-type-specific techniques like glutamate
uncaging (for example, see refs. 4–7), no tool has yet been invented
with the requisite spatiotemporal resolution to probe neural coding at
the resolution of single spikes. Furthermore, previous genetically
encoded optical methods, although elegant8–10,11, have allowed control
of neuronal activity over timescales of seconds to minutes, perhaps
owing to their mechanisms for effecting depolarization. Kinetics
roughly a thousand times faster would enable remote control of
Figure 1 ChR2 enables light-driven neuron
spiking. (a) Hippocampal neurons expressing
ChR2-YFP (scale bar 30 mm). (b) Left, inward
current in voltage-clamped neuron evoked by 1 s
of GFP-wavelength light (indicated by black bar);
right, population data (right; mean ± s.d. plotted
throughout; n ¼ 18). Inset, expanded initial
phase of the current transient. (c) Ten overlaid
current traces recorded from a hippocampal
neuron illuminated with pairs of 0.5-s light pulses
(indicated by gray bars), separated by intervals
varying from 1 to 10 s. (d) Voltage traces showing
membrane depolarization and spikes in a current-
clamped hippocampal neuron (left) evoked by
1-s periods of light (gray bar). Right, properties
of the first spike elicited (n ¼ 10): latency to
spike threshold, latency to spike peak, and
jitter of spike time. (e) Voltage traces in
response to brief light pulse series, with light
pulses (gray bars) lasting 5 ms (top), 10 ms
(middle) or 15 ms (bottom).
Published online 14 August 2005; doi:10.1038/nn1525
1Department of Bioengineering, Stanford University, 318 Campus Drive West, Stanford, California 94305, USA.2Max-Planck-Institute of Biophysics, Department of
Biophysical Chemistry, Max-von-Laue-Str. 3, D-60438 Frankfurt am Main, Germany.3Department of Biochemistry, Chemistry and Pharmaceutics, University of Frankfurt,
Marie-Curie-Str. 9, 60439 Frankfurt am Main, Germany.4Department of Psychiatry and Behavioral Sciences, Stanford School of Medicine, 401 Quarry Road, Stanford,
California 94305, USA.5Present address: Julius-von-Sachs-Institut, University of Wu ¨rzburg, Julius-von-Sachs-Platz 2–4, D-97082 Wu ¨rzburg, Germany. Correspondence
should be addressed to K.D. (firstname.lastname@example.org).
NATURE NEUROSCIENCE VOLUME 8 [ NUMBER 9 [ SEPTEMBER 20051263
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience
potassium gluconate, 38 KCl, 0.35 EGTA, 20 HEPES, 4 magnesium ATP, 0.35
sodium GTP, 6 NaCl, and 7 phosphocreatine (pH 7.25 with KOH). Neurons
were recorded in Tyrode solution (above). All experiments were performed at
room temperature (22–24 1C). For all experiments except for the synaptic
transmission data shown in Figure 4b,c, we patched fluorescent cells immersed
in Tyrode solution containing 5 mM NBQX and 20 mM gabazine to block
Photocurrents were measured while holding neurons in voltage clamp at
?65 mV. Recovery from inactivation was measured by measuring photo-
currents while illuminating neurons with pairs of 500-ms duration light pulses,
separated by periods of darkness lasting 1–10 s.
Spiking was measured while injecting current to keep the voltage of the cell
at approximately ?65 mV (holding current ranging from 0 pA to ?200 pA).
For synaptic transmission experiments, we patched nonfluorescent neurons
near ChR2-expressing neurons, immersed in Tyrode solution containing either
20 mM gabazine to isolate the excitatory postsynaptic response or in 5 mM
NBQX to isolate the inhibitory postsynaptic response. To confirm whether the
evoked potentials were indeed synaptically driven, after photostimulation, we
blocked all postsynaptic receptors with solution containing both 20 mM
gabazine and 5 mM NBQX and carried out photostimulation again.
pClamp 9 software (Axon Instruments) was used to record all data, and a
DG-4 high-speed optical switch with 300-W xenon lamp (Sutter Instruments)
was used to deliver the light pulses for ChR2 activation. An Endow GFP filter
set (excitation filter HQ470/40?, dichroic Q495LP; Chroma) was used for
delivering blue light for ChR2 activation. YFP was visualized with a standard
YFP filter set (excitation HQ500/20?, dichroic Q515LP, emission HQ535/30 m;
Chroma). Through a 20? objective lens, power density of the blue light was
8–12 mW/mm2, measured with a power meter (Newport).
Pulse series were synthesized by custom software written in MATLAB
(MathWorks) and then exported through pClamp 9 via a Digidata board
(Axon) attached to a PC. Poisson trains were generated in MATLAB by creating
series of pulses with inter-pulse intervals independently picked from a Poisson
distribution with mean l. Poisson trains were 8 s long, with mean interval
l ¼ 100 or 200 ms. For biophysical realism, a 10-ms minimum refractory
period was enforced between consecutive light pulses.
Membrane resistance was measured in voltage clamp mode with 20-mV
depolarizing steps lasting 75 ms. Spike rates due to direct current injection were
measured with 300-pA current steps lasting 0.5 s.
Data analysis. Data was analyzed using Clampfit (Axon) and custom software
written in MATLAB. Spikes were extracted by looking for voltage crossings of a
threshold (typically 60 mV above resting potential), and latencies were
measured from the onset of the light pulse to the spike peak. Extraneous
spikes were measured as the number of extra spikes after a single light pulse,
plus any spikes occurring later than 30 ms after the onset of a light pulse.
Jitter was calculated as the standard deviation of spike latencies, measured
either across all the spikes throughout a spike train (‘throughout-train’ jitter),
or for a particular spike across different trains (when gauging trial-to-trial
reliability, or across different neurons). For display of population data,
throughout-train jitter and trial-to-trial jitter were averaged across all neurons,
and neuron-to-neuron jitter was averaged across all spikes. For all jitter
analyses, light pulses which failed to elicit a spike were ignored. For the
across-neuron jitter analysis shown in Figure 2h, light pulses that did not elicit
spikes in all seven neurons were ignored (leaving 31/59 light pulses for the
l ¼ 100 ms stimulus, and 30/46 light pulses for the l ¼ 200 ms stimulus).
We would like to thank L. Meltzer and N. Adeishvili for experimental assistance;
C. Niell, C. Chan and J.P. Levy for helpful discussions and D. Ollig for technical
help. E.B. and G.N. are supported by the Max-Planck-Society and acknowledge
a grant from the German Research Foundation (DFG) in the research unit 472
(Molekulare Bioenergetik). E.S.B. is supported by the Helen Hay Whitney
Foundation, the Dan David Prize Foundation, and National Institute on
Deafness and Other Communication Disorders, and F.Z. is supported by a US
National Institutes of Health predoctoral fellowship. K.D. is supported by the
National Institute of Mental Health, the Stanford Department of Bioengineering,
the Stanford Department of Psychiatry and Behavioral Sciences, the Neuroscience
Institute at Stanford, the National Alliance for Research On Schizophrenia and
Depression and the Culpeper, Klingenstein, Whitehall, McKnight, and Albert Yu
and Mary Bechmann Foundations.
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Received 12 May; accepted 26 July 2005
Published online at http://www.nature.com/natureneuroscience/
1. Kandel, E.R., Spencer, W.A. & Brinley, F.J., Jr. Electrophysiology of hippocampal
neurons. I. Sequential invasion and synaptic organization. J. Neurophysiol. 24,
2. Ditterich, J., Mazurek, M.E. & Shadlen, M.N. Microstimulation of visual cortex affects
the speed of perceptual decisions. Nat. Neurosci. 6, 891–898 (2003).
3. Salzman,C.D.,Britten, K.H.& Newsome,
influences perceptual judgements of motion direction. Nature 346, 174–177
4. Shepherd, G.M., Pologruto, T.A. & Svoboda, K.Circuit analysis of experience-dependent
plasticity in the developing rat barrel cortex. Neuron 38, 277–289 (2003).
the intermediate gray layer of the superior colliculus. J. Neurophysiol. 81, 1424–1427
functional networks. Nature 433, 868–873 (2005).
7. Dalva, M.B. & Katz, L.C. Rearrangements of synaptic connections in visual cortex
revealed by laser photostimulation. Science 265, 255–258 (1994).
8. Lima, S.Q. & Miesenbock, G. Remote control of behavior through genetically targeted
photostimulation of neurons. Cell 121, 141–152 (2005).
9. Banghart, M., Borges, K., Isacoff, E., Trauner, D. & Kramer, R.H. Light-activated
ion channels for remote control of neuronal firing. Nat. Neurosci. 7, 1381–1386
10.Zemelman, B.V., Nesnas, N., Lee, G.A. & Miesenbock, G. Photochemical gating of
heterologous ion channels: remote control over genetically designated populations of
neurons. Proc. Natl. Acad. Sci. USA 100, 1352–1357 (2003).
11.Zemelman, B.V., Lee, G.A., Ng, M. & Miesenbock, G. Selective photostimulation of
genetically chARGed neurons. Neuron 33, 15–22 (2002).
12.Nagel, G. et al. Channelrhodopsin-2, a directly light-gated cation-selective membrane
channel. Proc. Natl. Acad. Sci. USA 100, 13940–13945 (2003).
13.Nagel, G. et al. Channelrhodopsin-1: a light-gated proton channel in green algae.
Science 296, 2395–2398 (2002).
14.Sineshchekov, O.A., Jung, K.H. & Spudich, J.L. Two rhodopsins mediate phototaxis to
low- and high-intensity light in Chlamydomonas reinhardtii. Proc. Natl. Acad. Sci. USA
99, 8689–8694 (2002).
15.Suzuki, T. et al. Archaeal-type rhodopsins in Chlamydomonas: model structure and
intracellular localization. Biochem. Biophys. Res. Commun. 301, 711–717 (2003).
16.Harz, H. & Hegemann, P. Rhodopsin-regulated calcium currents in Chlamydomonas.
Nature 351, 489–491 (1991).
268, 1503–1506 (1995).
18.Mermelstein, P.G., Bito, H., Deisseroth, K. & Tsien, R.W. Critical dependence of cAMP
responseelement-bindingprotein phosphorylation onL-typecalcium channels supports
a selective response to EPSPs in preference to action potentials. J. Neurosci. 20, 266–
19.Bi, G.Q. & Poo, M.M. Synaptic modifications in cultured hippocampal neurons:
dependence on spike timing, synaptic strength, and postsynaptic cell type. J. Neurosci.
18, 10464–10472 (1998).
20.Katz, L.C. & Dalva, M.B. Scanning laser photostimulation: a new approach for analyzing
brain circuits. J. Neurosci. Methods 54, 205–218 (1994).
21.Dantzker, J.L. & Callaway, E.M. Laminar sources of synaptic input to cortical inhibitory
interneurons and pyramidal neurons. Nat. Neurosci. 3, 701–707 (2000).
22.Schubert, D. et al. Layer-specific intracolumnar and transcolumnar functional connec-
tivity of layer V pyramidal cells in rat barrel cortex. J. Neurosci. 21, 3580–3592
J. Neurobiol. 51, 237–247 (2002).
24.Ikegaya, Y. et al. Synfire chains and corticalsongs: temporal modulesof cortical activity.
Science 304, 559–564 (2004).
25.Kozloski, J., Hamzei-Sichani, F. & Yuste, R. Stereotyped position of local synaptic
targets in neocortex. Science 293, 868–872 (2001).
26.Pouille, F. & Scanziani, M. Routing of spike series by dynamic circuits in the
hippocampus. Nature 429, 717–723 (2004).
27.Graziano, M.S., Taylor, C.S. & Moore, T. Complex movements evoked by microstimula-
tion of precentral cortex. Neuron 34, 841–851 (2002).
28.Moore, T. & Armstrong, K.M. Selective gating of visual signals by microstimulation of
frontal cortex. Nature 421, 370–373 (2003).
29.Deisseroth, K. et al. Excitation-neurogenesis coupling in adult neural stem/progenitor
cells. Neuron 42, 535–552 (2004).
30.Dull, T. et al. A third-generation lentivirus vector with a conditional packaging system.
J. Virol. 72, 8463–8471 (1998).
1268 VOLUME 8 [ NUMBER 9 [ SEPTEMBER 2005 NATURE NEUROSCIENCE
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience