of passive movement test performance.
Scatter plots reporting the detection angle and and plots reporting the error rate
(percentage correct trials ± 95% CI) on the TTDPM test performance without EES
and when delivering continuous EES at 0.8 and 1.5 times motor response threshold
amplitudes and a range of EES frequencies. Different EES frequencies were tested
on subject #1 (10 Hz, 50 Hz, 100 Hz) and subject #3 (30 Hz, 50 Hz). At 1.5 motor
response threshold amplitude, EES frequencies below 50 Hz induced spasms in the
muscles and were thus not tested. Grey dots report the detection angle for successful
trials, while pink dots and red crosses indicate false positive and failure to detect
movement within the allowed range of motion, respectively (n = 65 for subject #1
and n = 66 for subject #3). *, P < 0.05, Clopper-Pearson non-overlapping intervals,
two-sided.
2.
Supplementary Figure 2 Effect of EES on the natural modulation of
proprioceptive circuits during passive movements: extended data.
a, Configuration of the experimental setup for subject #1 and #3, as described in
Fig.
3a
. b, Plots showing EES pulses, EMG activity of the vastus medialis, and
changes in knee joint angle during passive oscillations of the knee when EES is
delivered at 60 Hz in subject #2 — similar results were achieved in subject #1 and
#3. Conventions as in Fig.
3b.
3.
Supplementary Figure 3 Impact of EES amplitude on muscle activity and leg
kinematics during locomotion on a treadmill: Subject #1.
a, AIS leg motor score. b, Configuration of electrodes targeting the left and right
posterior roots projecting to the L1 and L4 segments. Continuous EES was delivered
through these electrodes to facilitate locomotion. c, EMG activity of flexor
(semitendinosus/tibialis anterior) and extensor (rectus femoris/soleus) muscles
spanning the right knee and ankle joints, together with the changes in the knee ankle
and foot height trajectories over four gait cycles without EES and with EES
delivered at 0.9, 1.2 and 1.5 motor response threshold amplitude — similar results
were obtained for 30 gait cycles (analyzed in d). EES frequency was set to 40Hz. d,
Violin plots reporting the root mean square activity of the recorded muscles, the
range of motion of the knee and ankle angles, and the step height for different gait
cycles (n = 53 gait cycles). Small grey dots represent the different data points, while
the large white dots represent the median of the different distributions. Box and
whiskers report the interquartile range and the adjacent values, respectively. *, P <
0.05, ***, P < 0.001, Wilcoxon rank-sum two-sided test with Bonferroni correction
for multiple comparisons.
4.
Supplementary Figure 4 Impact of EES frequency and amplitude on muscle
activity and leg kinematics during locomotion on a treadmill: Subject #2.
The results displayed in Fig.
6
and Supplementary Figure
3
for subject #1 are
reported for subject #2 using the same conventions. Recordings in
panels a and c were repeated for 29 and 20 gait cycles and analyzed in
panels b and c, respectively. The statistics in panel d were computed over n = 37
gait cycles. *, P < 0.05, ***, P < 0.001, Wilcoxon rank-sum two-sided test with
Bonferroni correction for multiple comparisons.
activity and leg kinematics during locomotion on a treadmill: Subject #3.
The results displayed in Fig.
6
and Supplementary Figure
3
for subject #1 are
reported for subject #3 using the same conventions. Recordings in
panels a and c were repeated for 51 and 25 gait cycles and were analyzed in
panels b and c, respectively. The statistics in panel b and d were computed over n =
77 and n = 51 gait cycles, respectively. *, P < 0.05, ***, P < 0.001, Wilcoxon
rank-sum two-sided test with Bonferroni correction for multiple comparisons.
6.
Supplementary Figure 6 High-frequency, low-amplitude EES protocols preserve
proprioceptive information and promote motor patterns formation.
Impact of continuous high-frequency low-amplitude EES protocols (600 Hz, 20%
recruited afferents) on the modulation of the muscle spindle feedback circuits,
following the same conventions as in Fig.
5
. For comparison, the impact of
continuous EES on the group-Ia afferent firings is also reported.
7.
Supplementary Figure 7 Integrate-and-fire motor neuron model.
Schematic of the integrate and fire model and of the different synapses contacting
this cell. b, Simulated inhibitory and excitatory post synaptic potentials
(IPSPs/EPSPs) induced by the activation of a single Ia-inhibitory interneuron or a
single group-Ia afferent fiber, respectively. c, Excitation threshold of our
multicompartmental alpha motoneuron model. d, Number and amplitude of
experimental and modeled EPSP/IPSPs induced from the synaptic contacts
originating from group-Ia afferents (s1), group-II excitatory interneurons (s2), and
Ia-inhibitory interneurons (s3).
8.
Supplementary Figure 8 Adaptation of the rat neural network to humans.
a, Model layout of the hybrid rat-human computational model used to tune the
human neural network weights. W
1, w
2, w
3and w
4represent the weights of the neural
network connections that have been modified to adapt the rat neural network to the
human one. b, Systematic search results. W
1and w
3were ranged together between 1
and 2 times the weight used in the rat network, while w
2and w
4were ranged
between 1 and 4 times. Bar plots report the percentage of simulations that fulfilled
the defined fitness criteria. Selected weights that have been used for further
simulations are highlighted with an arrow. c, Effect of EES on the natural activity of
Ia-inhibitory interneurons and on the production of motor patterns during
locomotion, in the hybrid rat-human model for the selected set of synaptic weights.
Panels on the left report the average firing rate profiles of the Ia-inhibitory
interneuron populations associated to either the flexor or the extensor network, as
well as their modulation depth (mean ± SEM, n = 11 gait cycles). Similarly,
right-most panels represent the average firing rate profiles of motoneurons and their mean
firing rate activity during the phase in which they are active (mean ± SEM, n = 11
gait cycles). Effects of different EES frequencies and amplitudes are reported on the
top and bottom panels, respectively.
Impa Scatt perfo of EE moto dots move Pear act of continu
ter plots repor ormance witho ES frequencie or response th report the de ement within t rson non-overl uous EES on
rting the detec out EES and w
s. Different EE hreshold ampli etection angle the allowed ra apping interva the threshold
ction angle and when delivering
ES frequencie itude, EES fre e for successf ange of motio als, tow-sided.
d to detection
d and plots rep g continuous E s were tested equencies belo
ful trials, whil n, respectively
n of passive m
porting the erro EES at 0.8 an on subject #1 ow 50 Hz ind e pink dots a y (n = 65 for movement tes or rate (percen d 1.5 times m (10 Hz, 50 H uced spasms and red cross subject #1 an st performanc ntage correct t otor response z, 100 Hz) an in the muscle ses indicate fa nd n = 66 for ce. trials ± 95% C e threshold am d subject #3 ( es and were t alse positive a subject #3). * CI) on the TTDP mplitudes and a 30 Hz, 50 Hz) thus not tested
and failure to *, P < 0.05, C PM test a range ). At 1.5 d. Grey o detect
Clopper-Effec a, Co of th subje ct of EES on t onfiguration of e vastus med ect #2 — simil the natural m f the experime dialis, and cha
ar results were
modulation of
ental setup for anges in knee e achieved in proprioceptiv subject #1 and e joint angle d subject #1 and ve circuits du d #3, as descr during passive d #3. Convent uring passive ribed in Figure e oscillations o tions as in Fig movements e 3a. b, Plots of the knee w ure 3b. — extended d showing EES when EES is d data. pulses, EMG delivered at 60 activity 0 Hz in
Impa a, AI Cont anter ankle ampl the r differ medi ***, P act of EES am IS leg motor s tinuous EES w rior) and exten e and foot heig litude — simila root mean squ rent gait cycle ian of the diffe P < 0.001, Wilc mplitude on m score. b, Confi was delivered nsor (rectus fe ght trajectories ar results wer uare activity o es (n = 53 gait erent distributio coxon rank-su muscle activity iguration of ele d through thes emoris/soleus)
s over four gai re obtained fo of the recorde t cycles). Sma ons. Box and um two-sided t
y and leg kine
ectrodes targe se electrodes ) muscles span t cycles withou or 30 gait cycl ed muscles, th
all grey dots r whiskers repo test with Bonfe
ematics durin
eting the left a to facilitate l nning the righ ut EES and wi es (analyzed i he range of m represent the ort the interqua erroni correctio
ng locomotion
and right poste ocomotion. c, t knee and an ith EES delive in d). EES freq motion of the k different data artile range an on for multiple n on a treadm
erior roots proj , EMG activity nkle joints, tog ered at 0.9, 1.2 quency was s knee and ank points, while nd the adjacen e comparisons mill — Subject jecting to the L y of flexor (se ether with the 2 and 1.5 moto et to 40Hz. d, kle angles, and
the large whit nt values, resp . t #1. L1 and L4 seg emitendinosus changes in th or response th Violin plots re d the step he te dots repres pectively. *, P gments. s/tibialis he knee reshold eporting eight for sent the < 0.05,
Impa
The conv The Bonf
act of EES fre
results displa ventions. Reco statistics in pa ferroni correcti equency and a ayed in Figur ordings in pan anel d were c on for multiple amplitude on re 6 and Sup nels a and c w computed ove e comparisons n muscle activ pplementary were repeated er n = 37 gait s.
vity and leg k
Figure 3 for d for 29 and t cycles. *, P < kinematics du r subject #1 20 gait cycles < 0.05, ***, P ring locomot are reported s and analyze < 0.001, Wilco tion on a tread d for subject ed in panels b oxon rank-sum dmill — Subje #2 using the b and c, respe m two-sided te ect #2. e same ectively. est with
Impa
The conv respe 0.00
act of EES fre
results displa ventions. Reco ectively. The s 1, Wilcoxon ra equency and a ayed in Figur ordings in pa statistics in pa ank-sum two-s amplitude on re 6 and Sup anels a and c anel b and d sided test with
n muscle activ
pplementary c were repea
were comput Bonferroni co
vity and leg k
Figure 3 for ated for 51 a ted over n = orrection for m kinematics du r subject #1 and 25 gait c 77 and n = 5 ultiple compar ring locomot are reported cycles and we 51 gait cycles, risons. tion on a tread d for subject ere analyzed , respectively. dmill — Subje #3 using the in panels b *, P < 0.05, * ect #3. e same and c, ***, P <
High Impa spind Ia aff h-frequency lo act of continuo dle feedback c ferent firings is ow-amplitude ous high-frequ circuits, followi s also reported e EES protoco ency low-amp ing the same c d. ols preserve p plitude EES pr conventions a proprioceptiv rotocols (600 as in Figure 5. ve information Hz, 20% recru . For comparis n and promot uited afferents son, the impac
te motor patte s) on the mod ct of continuou erns formatio ulation of the us EES on the on. muscle e
group-Supp Integ Sche post fiber expe excit plementary F
grate and fire
ematic of the synaptic pote , respectively erimental and atory interneu igure 7 e motoneuron integrate and entials (IPSPs/ . c, Excitation modeled EP urons (s2), and n model. fire model a /EPSPs) induc n threshold o PSP/IPSPs ind d Ia-inhibitory nd of the diff ed by the act of our multico uced from th y interneurons ferent synapse tivation of a s ompartmenta he synaptic c s (s3). es contacting single Ia-inhib l alpha moto contacts origi this cell. b, S bitory interneu oneuron mode inating from imulated inhib uron or a sing el. d, Numbe group-Ia affe
bitory and exc gle group-Ia a er and amplit erents (s1), g citatory afferent tude of group-II
Supp Adap a, M w4 re hum netw plementary F ptation of the odel layout o epresent the an one. b, Sy work, while w2 igure 8 e rat neural n of the hybrid weights of t ystematic sea 2 and w4 were network to hu rat-human co he neural net rch results. W e ranged betw umans. omputational twork connec W1 and w3 we ween 1 and 4 model used t ctions that ha ere ranged to 4 times. Bar p to tune the h ave been mod ogether betwe plots report t uman neural dified to adap een 1 and 2 t he percentag network weig pt the rat ne times the we e of simulatio ghts. W1, w2, eural network ight used in ons that fulfil
w3 and
to the the rat led the
on the natural activity of Ia-inhibitory interneurons and on the production of motor patterns during locomotion, in the hybrid rat-human model for the selected set of synaptic weights. Panels on the left report the average firing rate profiles of the Ia-inhibitory interneuron populations associated to either the flexor or the extensor network, as well as their modulation depth (mean ± SEM, n = 11 gait cycles). Similarly, right-most panels represent the average firing rate profiles of motoneurons and their mean firing rate activity during the phase in which they are active (mean ± SEM, n = 11 gait cycles). Effects of different EES frequencies and amplitudes are reported on the top and bottom panels, respectively.
Gender Age (y) Years after SCI
WISCI II score AIS
Neurological level of injury UEMS total (max 50) LER motor subscore (max 25)
LEL motor subscore (max 25) LE motor score L2 (R/L) LE motor score L3 (R/L) LE motor score L4 (R/L) LE motor score L5 (R/L) LE motor score S1 (R/L) LE sensory score L1 (R/L) LE sensory score L2 (R/L) LE sensory score S1 (R/L) LE sensory score S2 (R/L) PPR sensory subscore PPL sensory subscore PPR, dermatome L1-S2 PPL, dermatome L1-S2 LE sensory score L3 (R/L) LE sensory score L4 (R/L) LE sensory score L5 (R/L) LTR sensory subscore LTL sensory subscore m m m 28 35 47 6 6 4 13 6 0 C D C* C7 C4 C7 46 31 45 14 13 0 0 12 0 2/0 2/2 0/0 2/0 4/4 0/0 4/0 2/1 0/0 4/0 1/1 0/0 2/0 4/4 0/0 38 29 26 37 36 29 17 29 13 16 36 15 0 4 0 0 8 0 1/1 1/1 0/0 1/1 1/1 0/0 1/1 1/1 0/1 1/1 1/1 0/1 1/1 0/1 0/1 1/1 0/1 1/1 1/1 1/2 0/0
AIS, American Spinal Injury Association Impairment Scale; LEMS, Lower extremity motor score; SCI, spinal cord injury; LEL, Lower extremity left; LER, lower extremity right; LTL, Light touch left; LTR, Light touch right; PPL, Pin prick left; PPR, Pin prick right; R/L, right/left; UEMS, Upper extremity motor score; WISCI, Walking index for spinal cord injury.
*Reason of AIS C classification in spite of motor scores of 0 throughout all lower extremity key muscles is the presence of voluntary anal contraction.
Supplementary Table 1 | Subjects’ data and neurological status according to the International Standards for Neurological Classification of Spinal Cord Injury. All the values were collected before the surgical