Spinal cord slice preparations. Experiments were performed on spinal sections isolated from 23 neonatal (P1-P16) Sim1 Cre/+ ;Rosa26 floxstopTdTom/+ /Gt(Rosa)26 floxstopH134R/EYFP/+ (referred to as Sim1TdTom/ChR2) mice. Isolation and preparation of sections, and electrophysiological recording methods have been previously described (Chopek, Nascimento et al. 2018) . Briefly, animals were anaesthetized with isoflurane, decapitated at the medulla-spinal cord junction and spinal cords dissected out in ice-cold dissecting artificial cerebrospinal fluid (aCSF), composed of (mM): KCl (3.5), NaHCO3 (35), KH2PO4 (1.2), MgSO4 (1.3), CaCl2 (1.2), glucose (10), sucrose (212.5), MgCl2 (2.2), and equilibrated to pH 7.4 with 95% 02 & 5% C02. Once dissected free, thoracic spinal cords were immediately secured in agarose and sectioned at 350 µm using a vibratome (Leica VT1200S, Leica) and then incubated in warm (30 o C) aCSF for a minimum of 30 minutes before performing electrophysiological experiments. Incubation and recording aCSF was composed of (mM): NaCl (111.0), KCl (3.085), D-glucose (10.99), NaHCO3 (25.0), MgSO4 . 7H2O (0.31), CaCl2 (2.52), KH2PO4
(1.1), equilibrated to pH 7.4 with 95% O2 & 5% CO2.
Intact spinal cord preparations. Spinal cords were dissected free from Sim1TdTom/ChR2 mice as described above in 'Slice preparations' but remained longitudinally intact with dorsal and ventral roots attached. Once free, connective tissue was carefully removed from spinal cord tissue and secured ventral side up with fine insect pins. To retrogradely label SPNs, rhodamine dextran amine dye (RDA) was applied to cut T6-T8 ventral roots on one or both sides of the spinal cord using glass suction pipettes with internal diameters of 100-120 µm. Ventral roots were cut close to their exit from the spinal cord to minimize labeling time. Retrograde labeling of thoracic SPNs continued in the dark at room temperature for at least 3 hours (Szokol, Glover et al. 2008) . A block of agar was prepared in advance with one side of the block cut with a scalpel to provide a 30-degree incline. The spinal cord was then mounted on the agar block and fixed in place with acrylic glue to expose the corresponding labeled thoracic spinal segment. The spinal cord at the level of the agar block was then sectioned with a vibratome. This portion of the spinal cord was then glued to a sylgard-coated (Sylgard, Dow Corning, MI, USA) recording chamber designed and 3D printed in-house specifically for these experiments. In particular, the obliquely cut surface of the spinal cord was placed on a sylgard 'ramp' to enable visualisation and patch-clamp recordings of SPNs under fluorescence while preserving and maintaining continuity with the lumbar spinal region for optical stimulation.
Whole cell patch-clamp recordings and optogenetic stimulation. Slices or spinal cord preparations were transferred to a recording chamber mounted on a Zeiss AxioExaminer microscope and perfused with oxygenated room-temperature aCSF. Cells were visualized using a 20x wide aperture (1.2 nA) water-immersion objective lens, a CCD camera (CoolSNAP EX OCD Camera, Photometrics, AZ) and Slidebook 6.0 software (Intelligent Imaging Innovations, CO, USA, RRID:SCR_014300). Patch pipettes were pulled with a P-97 Sutter puller and those with 4-6 MΩ resistances were filled with the following (mM): K-gluconate (128); NaCl (4); CaCl2 (0.0001); Hepes (10), glucose (1); Mg-ATP (5); and GTP-Li (0.3). Whole cell patch-clamp recordings were made under current-clamp using a Multiclamp 700B amplifier (Molecular Devices, California, USA, RRID: SCR_014300). Recordings were low pass filtered at10 kHz and acquired at 25 kHz with CED Power 1401 AD board and displayed and recorded using Signal software (Cambridge Electronic Design, Cambridge UK). Before performing an optical stimulation protocol, rheobase (defined as the minimum current to elicit a single AP was collected to determine cell excitability). Before the optical stimulation protocol, we recorded rheobase and repetitive firing in response to 1 second depolarizing current steps from each SPN. In slice preparations, presumed SPNs (small neurons visualized in the IML, with TdTom fluorescence noted near the soma) were patched and responses recorded. Using a spatial light modulator system as previously described (Chopek, Nascimento et al. 2018 , Chopek, Zhang et al. 2021 ), a region of interest slightly greater than the SPN soma was created for optical stimulation of presumed V3 terminals apposing the patched SPN. Blue light was delivered in five 500 ms pulses at 5 Hz at a laser power of ~2.5 mW. For intact cord preparations with thoracic surface exposed, SPNs that were retrogradely labelled with RDA were patched and responses recorded. A region of interest covering the ventral L2 segment was created for optical stimulation of presumed ventral L2 V3 neurons and likely axons of passage from caudal lumbar segments. Blue light was delivered in five 500 ms pulses at 5 Hz at a laser power of ~2.5 mW. In a subset of slice experiments, lumbar V3 neurons were patched and optically activated to confirm that optical stimulation resulted in AP generation in V3 neurons.
(1.1), equilibrated to pH 7.4 with 95% O2 & 5% CO2.
Intact spinal cord preparations. Spinal cords were dissected free from Sim1TdTom/ChR2 mice as described above in 'Slice preparations' but remained longitudinally intact with dorsal and ventral roots attached. Once free, connective tissue was carefully removed from spinal cord tissue and secured ventral side up with fine insect pins. To retrogradely label SPNs, rhodamine dextran amine dye (RDA) was applied to cut T6-T8 ventral roots on one or both sides of the spinal cord using glass suction pipettes with internal diameters of 100-120 µm. Ventral roots were cut close to their exit from the spinal cord to minimize labeling time. Retrograde labeling of thoracic SPNs continued in the dark at room temperature for at least 3 hours (Szokol, Glover et al. 2008) . A block of agar was prepared in advance with one side of the block cut with a scalpel to provide a 30-degree incline. The spinal cord was then mounted on the agar block and fixed in place with acrylic glue to expose the corresponding labeled thoracic spinal segment. The spinal cord at the level of the agar block was then sectioned with a vibratome. This portion of the spinal cord was then glued to a sylgard-coated (Sylgard, Dow Corning, MI, USA) recording chamber designed and 3D printed in-house specifically for these experiments. In particular, the obliquely cut surface of the spinal cord was placed on a sylgard 'ramp' to enable visualisation and patch-clamp recordings of SPNs under fluorescence while preserving and maintaining continuity with the lumbar spinal region for optical stimulation.
Whole cell patch-clamp recordings and optogenetic stimulation. Slices or spinal cord preparations were transferred to a recording chamber mounted on a Zeiss AxioExaminer microscope and perfused with oxygenated room-temperature aCSF. Cells were visualized using a 20x wide aperture (1.2 nA) water-immersion objective lens, a CCD camera (CoolSNAP EX OCD Camera, Photometrics, AZ) and Slidebook 6.0 software (Intelligent Imaging Innovations, CO, USA, RRID:SCR_014300). Patch pipettes were pulled with a P-97 Sutter puller and those with 4-6 MΩ resistances were filled with the following (mM): K-gluconate (128); NaCl (4); CaCl2 (0.0001); Hepes (10), glucose (1); Mg-ATP (5); and GTP-Li (0.3). Whole cell patch-clamp recordings were made under current-clamp using a Multiclamp 700B amplifier (Molecular Devices, California, USA, RRID: SCR_014300). Recordings were low pass filtered at10 kHz and acquired at 25 kHz with CED Power 1401 AD board and displayed and recorded using Signal software (Cambridge Electronic Design, Cambridge UK). Before performing an optical stimulation protocol, rheobase (defined as the minimum current to elicit a single AP was collected to determine cell excitability). Before the optical stimulation protocol, we recorded rheobase and repetitive firing in response to 1 second depolarizing current steps from each SPN. In slice preparations, presumed SPNs (small neurons visualized in the IML, with TdTom fluorescence noted near the soma) were patched and responses recorded. Using a spatial light modulator system as previously described (Chopek, Nascimento et al. 2018 , Chopek, Zhang et al. 2021 ), a region of interest slightly greater than the SPN soma was created for optical stimulation of presumed V3 terminals apposing the patched SPN. Blue light was delivered in five 500 ms pulses at 5 Hz at a laser power of ~2.5 mW. For intact cord preparations with thoracic surface exposed, SPNs that were retrogradely labelled with RDA were patched and responses recorded. A region of interest covering the ventral L2 segment was created for optical stimulation of presumed ventral L2 V3 neurons and likely axons of passage from caudal lumbar segments. Blue light was delivered in five 500 ms pulses at 5 Hz at a laser power of ~2.5 mW. In a subset of slice experiments, lumbar V3 neurons were patched and optically activated to confirm that optical stimulation resulted in AP generation in V3 neurons.
