SUPPLEMENTARY INFORMATION doi:10.

1038/nature11617

Materials and Methods

Subjects

Male Long-Evans rats weighing 200-225 grams (approximately 6-8 weeks) were obtained from
Charles River. Rats were maintained on a standard 12 hour light-dark cycle and given food and
water ad libitum. Rats were initially housed two per cage. Animals implanted with tetrode
microdrives or fixed wire arrays were housed individually after implantation to minimize damage to
recording hardware. Animals implanted with only fiber optics or cannulas continued to be housed
two per cage after implantation. All procedures conformed to guidelines established by the National
Institutes of Health and have been approved by the Stanford Institutional Animal Care and Use
Committee.

Automated Forced Swim Test (FST)

A 9-inch diameter tank of water (Tap Plastics, Mountain View, CA) was surrounded by a 10-inch
diameter coil constructed from 10 pounds of 26 gauge enamel-coated copper wire (Cal Coil
Magnetics, El Monte, CA). A 2g rare-earth magnet (Applied Magnets, Plano, TX) was placed on the
rat’s back foot with a comfortably snug rubber band. Magnet placement was performed immediately
before behavioral testing and was tolerated well by awake rats. During the FST, movement of the
magnet within the coil of wire during swimming was found to induce a robust current in the coil. Coil
voltage was bandpass filtered between 1 and 300 Hz, digitized at 2 kHz, and recorded for later
analysis using a Digital Lynx data acquisition system (Neuralynx, Bozeman, MT). Data was
simultaneously collected from a reference coil, and referencing was performed offline to reduce line
noise. The same system was used to measure locomotor activity in a familiar cage by placing the
induction coil directly underneath the cage; a similar method has been used previously to measure
Parkinsonian tremor in rats1. Coil data and video data were collected for all experiments. Manual
scoring for validation of the automatic method was performed blind to experimental condition and
automatically scored immobility/kick frequency. During manual scoring visual observations were
taken every 5 seconds, and immobile epochs were defined as either an absence of movement or
the minimum necessary movement required to stay afloat.

Tetrode microdrive and fixed wire array placement

W W W. N A T U R E . C O M / N A T U R E | 1
 RESEARCH SUPPLEMENTARY INFORMATION

For the data shown in Figure 2, rats reached a minimum of 400 g before surgery. For the data
shown in Supplementary Figures 6-9, rats were bilaterally injected with virus in the mPFC
(described below) at 8-10 weeks, and virus was allowed to express for a minimum of four months
before electrode implantation (rats typically reached weights > 400 g). Rats were initially
anesthetized with 5% isoflurane. The scalp was shaved and rats were placed in a stereotaxic frame
with non-rupturing ear bars. A heating pad was used to prevent hypothermia. Isoflurane was
delivered at 1-3% throughout surgery; this level was adjusted to maintain a constant surgical plane.
Ophthalmic ointment was used to protect the eyes. Buprenorphine (0.05 mg/kg, subcutaneous) and
enrofloxacin (5 mg/kg, subcutaneous) were given before the start of surgery. A mixture of 0.5%
lidocaine and 0.25% bupivicaine (100 µL) was injected subdermally along the incision line. The
scalp was disinfected with betadine and alcohol. A midline incision exposed the skull, which was
thoroughly cleaned. 8-11 skull screws were implanted at the periphery of the exposed skull to
ensure stable recordings, a 2-mm craniotomy was drilled over the right mPFC, and the dura mater
was carefully resected. A 4-tetrode microdrive (Neuralynx, Bozeman, MT) with tetrodes wound from
13 µm nichrome (Kanthal Palm Coast, Palm Coast, FL) or a 24-electrode fixed wire array with 50
µm stainless steel electrodes (NB Labs, Denison, TX) was implanted over the craniotomy (AP: 2.7
to 3.3 ML: 0.8 DV: 4.0). For the data shown in Supplementary Figures 6-9, a fiber optic targeting
the DRN was also implanted at this time (described below). Dental acrylic was used to secure the
electrodes and fiber to the skull and screws. The acrylic was shaped to make a thin neck between
the skull screws and the electrode interface board in order to facilitate waterproofing. The skin was
sutured closed and the rats were given carprofen (5 mg/kg, subcutaneous) and lactated ringer’s
solution (2.5 mL, subcutaneous) and recovered under a heat lamp. After implantation tetrodes were
adjusted daily.

Freely moving neurophysiology

Rats were briefly anesthetized with isoflurane. The headstage and tether (Neuralynx, Bozeman,
MT) were connected to the microdrive or fixed wire array and secured with thread to prevent
detaching during “wet dog” shakes. To protect the electronics from water damage a latex tube was
secured around the headstage attachment point with tightly wound rubber bands. At this time the
magnet was attached to the back foot for behavioral readout. The total time under isoflurane
anesthesia was limited to less than 10 minutes, and rats were allowed to recover for at least 1 hour
prior to the start of recordings. For the recordings in Supplementary Figures 6-9, the fiber optic
cable was attached immediately before the FST in order to minimize breakage from twisting during
rotation, and light power was checked immediately after recordings to confirm that the fiber optic
was intact.

2 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Neural data was acquired with a 64 channel Digital Lynx data acquisition system (Neuralynx,
Bozeman, MT). Spiking channels were first referenced to an electrode exhibiting no spiking activity
to reduce behavioral noise. The signal was then bandpass filtered between 600 and 6000 Hz and
digitized at 32 kHz. Induction coil data and video data were also recorded during all epochs in order
to validate the use of the induction coil method for both the FST and familiar cage activity. Data
was recorded for a variable number of epochs depending on the experiment. For Figure 2 we
recorded 15 minutes of data pre-FST in a familiar cage, 15 minutes during the FST, and 15 minutes
in a familiar cage post-FST. For Supplementary Figures 6-9 we recorded 15 minutes pre-FST in a
familiar cage, 20 minutes during the FST with stimulation (five two-minute no-stimulation epochs
interleaved with five two-minute stimulation epochs), 15 minutes in a familiar cage post-FST, and 20
minutes in a familiar cage post-FST with stimulation (five two-minute no-stimulation epochs
interleaved with five two-minute stimulation epochs). Rats were handled gently during transfer
between the familiar cage and the swim tank to minimize neural drift. After recording was
completed the waterproofing was removed and the rat was placed on a towel under a heat lamp for
10 minutes to dry. Before sacrifice for histology rats were deeply anesthetized and current passed
through all electrodes (50 µA for 30 seconds) to make electrolytic lesions for anatomical localization.
Data obtained using this method in freely moving rats is shown in Figures 1 and 2, and
Supplementary Figures 1-3 and 6-9.

Virus construction and packaging

Recombinant AAV vectors were serotyped with AAV5 coat proteins and packaged by the viral
vector core at the University of North Carolina. Viral titers were 2 x 1012 particles / mL for AAV5
CaMKII!::hChR2(H134R)-EYFP, 3 x 1012 particles / mL for AAV5 CaMKII!::EYFP, 4 x 1012 particles
/ mL for AAV5 CaMKII!::eNpHR3.0-EYFP, 4 x 1012 particles / mL for AAV5 hSyn::hChR2(H134R)-
EYFP, and 2 x 1012 particles / mL for AAV5 hSyn::EYFP. Maps are available online at
www.optogenetics.org.

Stereotaxic virus injection and optical fiber implantation

Rats were prepared for surgery and given analgesics and fluids as described above. A midline
incision exposed the skull, and craniotomies were made bilaterally above the mPFC. Virus was
injected with a 10 µL syringe and a 33 gauge beveled needle with the bevel facing anteriorly at 150
nL/min using an injection pump. Two 1 µL injections were delivered to each hemisphere at AP 2.2
mm, ML 0.5, DV 5.2 and AP 2.2, ML 0.5, DV 4.2 for a total of 4 µL per rat. After each injection the

W W W. N A T U R E . C O M / N A T U R E | 3
 RESEARCH SUPPLEMENTARY INFORMATION

needle was left in place for 7 minutes and then slowly withdrawn. The skin was sutured closed. For
the hSyn::ChR2 DRN rats a similar protocol was followed, but only one injection of 1 µL was
delivered to the DRN from the right at an angle of 20 degrees with the bevel facing medially.
Coordinates for DRN injections were AP -7.8 mm, ML 0.0, DV: 6.7. In projection-targeting
experiments virus was allowed to express for a minimum of 4 months in order to allow time for
sufficient opsin accumulation in the axons, while cell body stimulation experiments were conducted
at 8 weeks.

At least 10 days before behavioral testing a fiber optic with an external metal ferrule (200 µm
diameter, 0.22 NA, Doric Lenses, Québec, Canada) or a 22GA cannula (Plastics One, Roanoke,
VA) was implanted over the target structure of interest, as described previously 2. Coordinates for
mPFC implantation were AP 2.7, ML 0.5, DV 3.8. DRN fibers were implanted at a 30º angle from
the right to avoid both the central sinus and the cerebral aqueduct, and the coordinates for the tip of
the fiber were AP -7.8, ML 0.5, DV 5.9. When cannulas were used in the DRN for the glutamate
antagonist experiments they were also implanted at a 30º angle from the right, and the coordinates
at the tip were AP: -7.8, ML: 1.0, DV: 4.67. Internal cannulas for drug infusion extended 2 mm past
the tip of the cannula (to the center of the DRN), while the inserted fiber extended 1 mm past the tip
of the cannula in order to provide illumination to the whole DRN. LHb fiber optics were implanted
bilaterally at a 10º angle from midline on each side at AP: -3.6, ML: +/-0.8, DV: 4.6. Rats were
prepared for surgery and given analgesics and fluids as described above. A midline incision was
made, the skull was thoroughly cleaned, and a craniotomy was made over the target structure of
interest. Four to six skull screws were attached, and the fiber optic was lowered. A thin layer of
metabond (Parkell, Inc., Edgewood, NY) was used to firmly attach the hardware to the skull, which
was followed by a thicker layer of dental acrylic for structural support.

Light delivery

During behavioral testing an external optical fiber (200 µm diameter, 0.22 NA, Doric Lenses,
Québec, Canada) was coupled to the implanted fiber optic with a zirconia sleeve. If cannulas were
used, as for the glutamate antagonist experiments, a fiber for insertion through the cannula was
constructed from the same optical fiber coupled to an internal cannula (Plastics One, Roanoke, VA).
An optical commutator allowed for unrestricted rotation (Doric Lenses, Québec, Canada)3. Optical
stimulation was provided with a 100 mW 473 nm or 594 nm diode pumped solid state laser (OEM
Laser Systems, Inc., Salt Lake City, UT) and controlled by a Master-8 stimulus generator (A.M.P.I.,
Jerusalem, Israel). Light pulses were recorded with a Digital Lynx data acquisition system
(Neuralynx, Bozeman, CA) simultaneously with behavioral and neural data. Pulse trains with 5 ms

4 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

long 473 nm light pulses at 20 Hz were used for all ChR2 experiments, while continuous 5 mW 594
nm light was used for all eNpHR3.0 experiments. The mPFC cell body stimulation experiments
used 3 mW light (24 mW/mm2 at the fiber tip); higher light power induced seizures. The mPFC-
DRN axonal stimulation experiments used 10-20 mW light (79-159 mW/mm2 at the fiber tip).
Greater light power was required during the DRN axonal stimulation experiments because of the
lower EYFP fluorescence at this site, a predictor of lower opsin expression. The mPFC-LHb and
DRN cell body experiments used 10 mW light (79 mW/mm 2 at the fiber tip).

In Vivo Pharmacology

A glutamate antagonist cocktail of either 22 mM NBQX and 38 mM AP5 or 8 mM NBQX and 13 mM

AP5 was used; both concentrations were effective at blocking the stimulation-induced behavioral
effect, and therefore the data from both concentrations was pooled. Ten minutes before the onset
of behavioral testing, 0.4 "L of this solution was infused through an internal cannula into the center
of the DRN. Solution was infused at a rate of 0.1 "L/min through a 28GA internal cannula (Plastics
One, Roanoke, VA) extending 2 mm past the end of the implanted cannula. A 10"L NanoFil syringe
(WPI, Sarasota, FL) was used in a syringe pump (Harvard Apparatus, Holliston, MA). The internal
cannula was left in place for 2 minutes after infusion, and then gently withdrawn and replaced with a
fiber optic extending 1 mm past the end of the implanted cannula.

Forced Swim Test

We utilized the Porsolt Forced Swim Test for these experiments4. The swim tank was filled with
25°C water to a height of 40 cm. The induction coil was placed around the tank at the level of the
feet, and a small magnet was attached to the rat’s back foot, as described above. The rats were
placed in the FST for 15 minutes on the first day during the light part of the light/dark cycle for pre-
exposure. They were then dried with a towel and placed under a heat lamp for 10 minutes to warm
before returning to the home cage. Data was collected during a second 15-20 minute test
performed 24 hours later. An external fiber optic was suspended above the FST tank and attached
to the implanted fiber optic with a zirconia sheath. During experiments with light stimulation,
stimulation alternated between on and off in two minute epochs, starting with no stimulation, using
the parameters described above for a total of 20 minutes. Induction coil data, video data, and laser
pulse time data were collected for all experiments. The FST tank water was changed between each
animal.

W W W. N A T U R E . C O M / N A T U R E | 5
 RESEARCH SUPPLEMENTARY INFORMATION

Open Field Test

An external fiber optic was suspended above the open field and attached to the implanted fiber optic
with a zirconia sheath. Rats were placed in the center of a white, dimly lit open field chamber (105 x
105 cm) and allowed to freely explore the environment. Light stimulation alternated between on and
off in three minute epochs, starting with no stimulation, for a total of 15 minutes. A video camera
was placed directly above the open field, and locomotor activity was detected and analyzed with
Viewer2 software (BiObserve, Fort Lee, NJ). Laser pulse time data was collected and synchronized
to behavioral data.

Anesthetized in vivo recordings

To confirm function expression of opsin, simultaneous dual site recording and optical stimulation of
the mPFC and the DRN was performed as described previously3 in anesthetized rats transduced in
the mPFC with the AAV5 CaMKII!::ChR2-EYFP construct. Data obtained with this method are
shown in Supplementary Figures 4 and 5. Rats were deeply anesthetized with isoflurane before the
start of recording. A midline incision was made and the skin reflected. 2-3 mm diameter
craniotomies were made above the mPFC (vertical penetration) and the DRN (30° penetration). A 1
Mohm epoxy-coated tungsten electrode (A-M Systems, Sequim, WA) coupled to a 200 µm 0.37 NA
optical fiber (Thorlabs Inc., Newton, NJ) was stereotaxically lowered until a unit was isolated starting
at AP 2.7, ML 0.5, DV 3.6 for the mPFC recordings, and AP -7.8, ML 0.5, DV 6.0 (30° penetration)
for the DRN recordings. Recorded signals were bandpass filtered between 0.3 and 10 kHz,
amplified 10000x (A-M Systems), digitized at 30 kHz (Molecular Devices, Sunnyvale, CA) and
recorded with Clampex software (Molecular Devices). Optical stimulation was provided with a 100
mW 473 nm diode pumped solid state laser (OEM Laser Systems, Inc., Salt Lake City, UT).
Clampex software was used for both recording neural data and controlling laser output. Light
2 2
powers between 1 mW (8 mW/mm at the fiber tip) and 20 mW (159 mW/mm ) were used. At the
end of all experiments current was passed through the electrode (50 µA for 30 seconds) to make an
electrolytic lesion for anatomical localization.

Histology, immunohistochemistry, and confocal imaging

6 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Rats were deeply anesthetized with Beuthanasia-D and transcardially perfused with ice-cold 4%
paraformaldehyde (PFA) in PBS (pH 7.4). Brains were fixed in PFA overnight and then transferred
to 30% sucrose in PBS to equilibrate for at least 3 days. 40 µm coronal sections were cut on a
freezing microtome and stored in cryoprotectant at 4°C. Sections were washed with PBS and
incubated for 30 minutes in PBS++ (PBS with 0.3% Triton X-100 and 3% normal donkey serum
(NDS)). Sections were incubated with primary antibody overnight in PBS++ at 4°C. Primary
antibodies used were rabbit anti-5-HT 1:1000 (ImmunoStar 20080, Hudson, WI) and rabbit anti-
GABA 1:500 (Sigma A2052, St. Louis, MO). Sections were then washed in PBS and incubated with
secondary antibody conjugated to Cy3 or Cy5 for three hours at room temperature in PBS++ (1:500,
Jackson Laboratories, West Grove, PA). Sections were washed in PBS and incubated in 4#,6-
diamidino-2-phenylindole (DAPI, 1:50000) for 10 minutes, then washed again and mounted on
slides with PVA-DABCO. Images were acquired using a Leica TCS SP5 confocal scanning laser
microscope with a 10X air objective or a 20X or 40X oil immersion objective.

Data Analysis

Spikes were imported into Offline Sorter software (Plexon Inc, Dallas, TX) and sorted offline using
waveform features (peak and valley heights) and principal components. Analyses of neural data
and behavioral data were done using custom written Matlab (Mathworks, Natick, MA) scripts and
the Neuroexplorer data analysis package (Nex Technologies, Littleton, MA). Statistical significance
was defined as p <= 0.05 for all analyses.

Induction coil data was referenced and zero-phase filtered offline at 1-6 Hz, which preserved the
shape of individual kicks while reducing high frequency line noise. The referenced and filtered coil
data was then integrated and thresholded (at 10% of the maximum deviation) and the peaks
corresponding to individual kicks were detected (Supplementary Figure 1). Kicks made during
struggling corresponded to large deflections in the recorded signal and could easily be separated
from periods of passive floating. Instantaneous kick frequency was defined as the average number
of kicks per second in 10 second bins. Automatically scored immobility was determined by scoring
epochs with a gap between kicks greater than one second as immobile. The same analysis was
performed on induction coil data collected during familiar cage recordings. In this case, the
induction coil data was filtered at 1-20 Hz, thresholded at 4% of the maximum deviation, and was
not integrated because the unipolar nature of the waveform. Automatically scored behavioral data
was regressed against manually scored behavioral data to determine the correspondence between
scoring methods. R-square and F statistics and p values were derived from this regression
analysis.

W W W. N A T U R E . C O M / N A T U R E | 7
 RESEARCH SUPPLEMENTARY INFORMATION

Determination of statistically significant differences in neural firing rate between different behavioral
epochs was done using the Mann-Whitney U test. Neurons were first tested for differences in firing
rate between the pre-FST epoch and the FST epoch. For this analysis neural and behavioral data
was binned in 10-second intervals. Neurons were then tested for differences in firing rate between
mobile and immobile states during the FST. For this analysis, mobile and immobile behavioral
epochs were divided into two different "continuous" data streams and then statistically tested as
above using 10-second bins. The selectivity index used in Figure 2 was defined as the difference in
firing rate between conditions divided by the sum. Criteria for identifying putative fast spiking
inhibitory neurons were a mean firing rate over 20 Hz and a narrow waveform5.

Statistical significance of the behavioral data in Figures 3 and 4 was determined using the Wilcoxon
signed-rank test. Data was first exponentially detrended. This was necessary because a decrease
in struggling with time upon exposure to the forced swim test environment is well known and has
been previously described67. Similarly, a decrease in locomotor activity with time upon exposure to
the open field test is usually reported8.

The instantaneous average firing rate depicted in Supplementary Figure 6 was calculated in 10-
second bins, and statistical significance for individual neurons was again calculated using the Mann-
Whitney U test. The distribution of mobility-immobility differences was tested for changes in
variance between stimulated and non-stimulated conditions using the F test for equal variances.
Differences in slope were tested with analysis of covariance (ANCOVA).

For the latency analysis in Supplementary Figure 3 a "baseline" firing rate was first established for
each neuron using the -15 to -5 seconds before the onset of mobility. A PSTH was made from the
data in this time period with 50 ms bins, smoothed with a 150 ms gaussian kernel, and the mean
and standard deviation were calculated across bins. A second PSTH was then made for the 10
seconds surrounding the onset of mobility (-5 to +5), and also used 50 ms bins and was smoothed
with the same kernel. The latency to activation was defined as the first bin of the first sequence
during this epoch of at least 10 bins where all firing rates were above the mean plus 2.58 times the
standard deviation. Relatively large bins were used for this analysis due to the typically low firing
rates of mPFC neurons and the low number of immobile-to-mobile transitions.

8 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 1: Detection of individual kicks in the Forced Swim Test. a) The
induction coil trace was first zero-phase filtered (1-6 Hz) and then integrated to yield a peak at the
midpoint of each kick. The integrated trace was then thresholded (10% of the maximum deviation)
and the peaks were detected. The threshold is shown in gray. b) The filtered coil trace before
integration. Kick times correspond to the midpoint of each kick.

W W W. N A T U R E . C O M / N A T U R E | 9
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 2: The magnetic induction method can be used to detect immobility
in a cage. a) The induction coil trace was zero-phase filtered (1-20 Hz), thresholded (4% of the
maximum deviation), and the peaks were detected. The cage coil trace was not integrated before
peak detection because of the unipolar waveform associated with steps. b) Automatically scored
cage immobility corresponds well to manually scored cage immobility. c) Average step frequency
corresponds well to manually scored immobility.

1 0 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 3: mPFC neurons activated during mobile epochs show elevated
firing preceding the onset of mobility epochs. a) Peristimulus time histogram (PSTH) for an
example mPFC neuron recorded during the FST. The latency to activation (see Supplementary
Methods) for this neuron preceded the onset of mobility by 0.4 s. Solid vertical line indicates
latency. b) Latency distribution for the population of mPFC neurons activated during mobile
epochs.

W W W. N A T U R E . C O M / N A T U R E | 1 1
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 4: mPFC functional ChR2 expression and open field test a) AAV5
CaMKII!::ChR2-EYFP was infused bilaterally in the mPFC. An optrode recording in an
anesthetized rat detected spiking activity in the mPFC induced by local cell body illumination. b)
Open field test. AAV5 CaMKII!::ChR2-EYFP or AAV5 CaMKII!::EYFP was infused bilaterally in the
mPFC. Light stimulation of the mPFC did not affect velocity in either ChR2-EYFP rats (Wilcoxon
signed-rank test, p=0.50, n=10) or EYFP rats (Wilcoxon signed-rank test, p=0.09, n=8). Red line
indicates the ChR2-EYFP group average. Gray line indicates the EYFP group average. Blue bars
indicate light on. Significance calculations were performed on detrended data.

1 2 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 5: DRN histology and optrode recording. a) AAV5 CaMKII!::ChR2-

EYFP was infused bilaterally in the mPFC. EYFP fluorescence in mPFC axons in the DRN is
shown in green and immunostaining for 5-HT is shown in red. b) AAV5 CaMKII!::ChR2-EYFP was
infused bilaterally in the mPFC. To confirm functional opsin expression in DRN-projecting mPFC
axons, an anesthetized optrode recording was performed in the DRN. Local spiking activity in the
DRN induced by illumination of mPFC axons in the DRN is shown. Spikes were not elicited with
every light pulse. 12 overlaid traces are shown.

W W W. N A T U R E . C O M / N A T U R E | 1 3
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 6: Optogenetic stimulation of DRN-projecting mPFC neurons

decreases mPFC encoding of mobility. a) AAV5 CaMKII!::ChR2-EYFP was infused bilaterally
into the mPFC, and a fiber optic was implanted over the DRN. A 24-electrode fixed-wire array was
implanted over the mPFC. b) Injected and implanted rats (n=3) showed a robust increase in FST
mobility during stimulation. c) Light stimulation influenced firing rate in almost all recorded mPFC
neurons (31/34, 91%, Mann-Whitney U test). The average firing rate across the population
increased slightly during stimulation epochs, although more neurons were significantly inhibited
(22/31, 71%) than excited (9/31, 29%) by stimulation. d) Illumination of mPFC axons in the DRN
decreased mPFC encoding of mobility state during the FST. Each point represents one neuron.
Black squares depict mobile vs. immobile firing rate of neurons without light stimulation, while blue
circles depict mobile vs. immobile firing rates with light stimulation. Light stimulation causes these
points to cluster tightly around the best-fit line. During epochs without light stimulation, 62% (21/34)
of neurons were significantly modulated by mobility state, but when these same neurons were
tested during light stimulation, the significantly selective proportion fell to 21% (7/34). There was
not a significant change in slope with light stimulation (ANCOVA, p=0.20). e-f) Histograms of the
change in firing rate between mobile and immobile states. Top: no light stimulation. Bottom: light
stimulation. Illumination decreases the variance in this distribution (F test for equal variance,
p=2.11e-4), indicating decreased encoding of mobility state in these neurons. g) All four quadrants
of neurons (see Fig. 2g) show a decrease in encoding of mobility state with light stimulation. h-j)
Stimulation does not have a significant effect on mobility state encoding during the post-FST epoch.
Neurons are the same as those shown in Supplementary Figures 7, 8, and 9.

1 4 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 7: Light onset-triggered mPFC PSTHs. Peristimulus time histograms of

34 mPFC neurons recorded during optical stimulation of mPFC axons in the DRN during the FST.
Histograms are aligned to the start of two-minute light-on stimulation epochs. Neurons are the
same as those shown in Supplementary Figures 6, 8, and 9.

W W W. N A T U R E . C O M / N A T U R E | 1 5
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 8: Light offset-triggered mPFC PSTHs. Peristimulus time histograms of

34 mPFC neurons recorded during optical stimulation of mPFC axons in the DRN during the FST.
Histograms are aligned to the end of two-minute light-on stimulation epochs. Neurons are the same
as those shown in Supplementary Figures 6, 7, and 9.

1 6 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 9: Perievent rasters for all mPFC neurons recorded during optical
stimulation of the mPFC-DRN projection. a) All spikes for all recorded neurons aligned to the
start of stimulation. Each shaded horizontal bar indicates 1 neuron with 5 stimulus onset
repetitions. Black ticks indicate spikes; vertical blue bars indicate light pulses. b) 33/34 neurons did
not exhibit a low-latency, low-jitter response to stimulation as shown in part a, but 1/34 neurons
(neuron 15) did. However, stimulation-driven spiking in this neuron was unreliable during the first
minute of stimulation, and this neuron did not pass the collision test, ruling out the possibility of
direct antidromic activation. Blue ticks indicate light pulses, black ticks indicate action potentials.
Neurons are the same as those shown in Supplementary Figures 6, 7, and 8.

W W W. N A T U R E . C O M / N A T U R E | 1 7
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 10: Glutamate antagonists in the DRN block the effect of mPFC-DRN
stimulation. a) Rats were infused bilaterally in the mPFC with AAV5 CaMKII!::ChR2-EYFP and
implanted with a cannula targeting the DRN. Six rats were tested twice in the FST
(counterbalanced), each time receiving either an infusion of saline or a cocktail of the glutamate
antagonists NBQX and AP5 10 minutes before testing. b) On average, rats receiving intra-DRN
NBQX+AP5 were significantly more mobile in the FST (p=3.73e-8, Mann-Whitney U test) than rats
receiving saline. c) 3/6 rats did not show a within-rat significant (p>0.05, Mann-Whitney U test)
change in mobility, and were used for the analysis in panels e and f. Each line represents one rat.
d) Application of blue light to ChR2-expressing axons in the DRN induced a robust behavioral
activation during the FST when saline was infused (red line, n=6); this activation was blocked by
infusion of NBQX + AP5 in the DRN (green line, n=6). All rats were included in this panel, therefore
the drug-induced increase in mobility is also evident. Blue bars indicate stimulation epochs. e-f)
When rats with differences in mobility between saline and NBQX+AP5 were excluded, rats infused
with saline continued to exhibit behavioral activation upon light exposure (Wilcoxon signed-rank
test, p=3.05e-4), and NBQX+AP5 continued to block the stimulation-induced increase in FST
mobility (Wilcoxon signed-rank test, p=0.93).

1 8 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 11: Optogenetic inhibition of DRN-projecting mPFC axons. a) Rats

were infused bilaterally in the mPFC with either AAV5 CaMKII!::eNpHR3.0-EYFP or AAV5
CaMKII!::EYFP and implanted with a fiber optic targeting the DRN. b) EYFP fluorescence in the
DRN. c) Forced swim test behavioral data from eNpHR3.0-expressing (n=9, red line) and EYFP-
expressing (n=10, black line) rats. Yellow bars indicate light on. Each epoch is 2 minutes. d) Kick
frequency during the first light-off epoch was not distinguishable between conditions (p=0.55, Mann-
Whitney U test). e) Consistent with the time course suggested in (c), kick frequency pooled over all
light-on epochs or all post-first-light-exposure epochs was significantly different between the
eNpHR3.0 and EYFP groups (p=0.0013 and p=0. 7.76e-7 respectively, Mann-Whitney U test).

W W W. N A T U R E . C O M / N A T U R E | 1 9
 RESEARCH SUPPLEMENTARY INFORMATION

Supplementary Figure 12: AAV5 hSyn::ChR2-EYFP is expressed in both 5-HT and GABA
DRN neurons. a) AAV5 hSyn::ChR2-EYFP was infused in the DRN. EYFP fluorescence in DRN
neuronal cell bodies is shown in green, immunostaining for 5-HT is shown in red, and DAPI staining
for nuclei is shown in white. b) EYFP fluorescence in DRN cell bodies is shown in green,
immunostaining for GABA is shown in blue, and DAPI staining for nuclei is shown in white.

2 0 | W W W. N A T U R E . C O M / N A T U R E
 SUPPLEMENTARY INFORMATION RESEARCH

Supplementary Figure 13: Stimulation of the mPFC-LHb projection affected locomotor

activity in the open field test. AAV5 CaMKII!::ChR2-EYFP or AAV5 CaMKII!::EYFP was infused
in the mPFC and bilateral fiber optics were implanted over the LHb. Light stimulation of the mPFC-
LHb projection significantly affected velocity in ChR2-EYFP rats (Wilcoxon signed-rank test,
p=0.039, n=4) but not EYFP rats (Wilcoxon signed-rank test, p=0.157, n=9). Red line indicates the
ChR2-EYFP group average. Gray line indicates the EYFP group average. Blue bars indicate light
on. Significance calculations were performed on detrended data.

W W W. N A T U R E . C O M / N A T U R E | 2 1
 RESEARCH SUPPLEMENTARY INFORMATION

References

1. Dill, R. E., Dorman, H. L. & Nickey, W. M. A simple method for recording tremors in small
animals. Journal of applied physiology 24, 598–9 (1968).

2. Zhang, F. et al. Optogenetic interrogation of neural circuits: technology for probing

mammalian brain structures. Nature protocols 5, 439–56 (2010).

3. Gradinaru, V. et al. Targeting and readout strategies for fast optical neural control in vitro and
in vivo. Journal of neuroscience 27, 14231–8 (2007).

4. Porsolt, R., Le Pichon, M. & Jalfre, M. Depression: a new animal model sensitive to
antidepressant treatments. Nature 266, 730–732 (1977).

5. Diester, I. & Nieder, A. Complementary contributions of prefrontal neuron classes in abstract

numerical categorization. Journal of neuroscience 28, 7737–47 (2008).

6. Hédou, G., Pryce, C., Di Iorio, L., Heidbreder, C. a & Feldon, J. An automated analysis of rat
behavior in the forced swim test. Pharmacology, biochemistry, and behavior 70, 65–76
(2001).

7. Cryan, J. F., Markou, A. & Lucki, I. Assessing antidepressant activity in rodents: recent
developments and future needs. Trends in pharmacological sciences 23, 238–245 (2002).

8. Walsh, R. N. & Cummins, R. a The Open-Field Test: a critical review. Psychological bulletin
83, 482–504 (1976).

2 2 | W W W. N A T U R E . C O M / N A T U R E