Neuroscience 224 (2012) 48–62

EFFECTS OF TRUNK-TO-HEAD ROTATION ON THE LABYRINTHINERESPONSES OF RAT RETICULAR NEURONS

M. BARRESI, a C. GRASSO, a L. BRUSCHINI, b

S. BERRETTINI b AND D. MANZONI a*aDepartment of Physiological Sciences, Pisa University,

I-56127 Pisa, Italy

b II� ENT Department, Pisa Hospital, I-56127 Pisa, Italy

Abstract—Vestibulospinal reflexes elicited by head dis-

placement become appropriate for body stabilization owing

to the integration of neck input by the cerebellar anterior ver-

mis. Due to this integration, the preferred direction of spinal

motoneurons’ responses to animal tilt rotates by the same

angle and by the same direction as the head over the body,

which makes it dependent on the direction of body displace-

ment rather than on head displacement. It is known that the

cerebellar control of spinal motoneurons involves the retic-

ular formation. Since the preferred directions of corticocere-

bellar units’ responses to animal tilt are tuned by neck

rotation, as occuring in spinal motoneurons, we investi-

gated whether a similar tuning can be observed also in the

intermediate station of reticular formation. In anaesthetized

rats, the activity of neurons in the medullary reticular forma-

tion was recorded during wobble of the whole animal at

0.156 Hz, a stimulus that tilted the animal’s head by a con-

stant amplitude (5�), in a direction rotating clockwise or

counter clockwise over the horizontal plane. The response

gain and the direction of tilt eliciting the maximal activity

were evaluated with the head and body axes aligned and

during a maintained body-to-head displacement of 5–20�over the horizontal plane, in either direction. We found that

the neck displacement modified the response gain and/or

the average activity of most of the responsive neurons.

Rotation of the response direction was observed only in a

minor percentage of the recorded neurons. The modifica-

tions of reticular neurons’ responses were different from

those observed in the P-cells of the cerebellar anterior ver-

mis, which rarely showed gain and activity changes and

often exhibited a rotation of their response directions. In

conclusion, reticular neurons take part in the neck tuning

of vestibulospinal reflexes by transforming a head-driven

sensory input into a body-centred postural response. The

present findings prompt re-evaluation of the role played by

0306-4522/12 $36.00 � 2012 IBRO. Published by Elsevier Ltd. All rights reservehttp://dx.doi.org/10.1016/j.neuroscience.2012.08.011

*Corresponding author. Address: Dipartimento di Scienze Fisiologi-che, Universita di Pisa, Via S. Zeno 31, 56127 Pisa, Italy.Tel: +39-50-2213466; fax: +39-50-2213527.

E-mail address: manzoni@dfb.unipi.it (D. Manzoni).Abbreviations: BF, base frequency; CCW, counter clockwise; CW,clockwise; D, direction; ND, nose-down; NU, nose-up; RF, reticularformation; SD, side-down; Smax, maximal sensitivity vector; Smin,minimal sensitivity vector; SN, signal-to-noise; SPDH, sequential pulsedensity histogram; STC, spatiotemporal convergence; SU, side-up; VS,vestibulospinal.

48

the reticular neurons and the cerebellum in vestibulospinal

reflexes. � 2012 IBRO. Published by Elsevier Ltd. All rights

reserved.

Key words: vestibulospinal reflexes, neck input, cerebellum,

reticular formation.

INTRODUCTION

Vestibular information elicited by head displacement in

space exerts a prominent role in the control of posture

(Magnus, 1928; Roberts, 1978), particularly after

reduction of visual and proprioceptive information or

when the base of the support is unstable (Igarashi et al.,

1970; Lackner et al., 1999; Welgampola and Colebatch,

2001; Cenciarini and Peterka, 2006). Labyrinthine

signals elicit vestibulospinal (VS) reflexes that are

spatially organized, each muscle being maximally

activated for a particular (preferred) direction of head

displacement (Wilson et al., 1986). For instance, in the

decerebrate cat, the maximal activation of the forelimb

extensor triceps brachii takes place during a roll tilt in

the frontal plane directed towards the recording side.

However, vestibular signals must be integrated with

proprioceptive neck signals related to the body-to-head

position (von Holst and Mittelstaedt, 1950; Roberts,

1978) in order to stabilize the body position. This

process has two different aspects: one is that the

vestibular and neck reflexes elicited by coplanar head

and neck rotation interact in order to modify the postural

tone only when the position of the trunk in space

changes, as pointed out by von Holst and Mittelstaedt

(1950) and Roberts (1978). It is well established that

this interaction of VS and cervicospinal reflexes is nearly

linear (Lindsay et al., 1976; Ezure and Wilson, 1983;

Manzoni et al., 1983). This mechanism, which has been

documented in decerebrate preparations, also underlies

the perception of body motion in healthy humans

(Mergner et al., 1991, 1997).

The other aspect, which is addressed in the present

study, is that a maintained head-to-body orientation,

which elicits a tonic neck input, allows one to infer body

motion from labyrinthine signals, leading to substantial

changes in the pattern of VS reflexes. As shown in

Fig. 1, the same labyrinthine signals are elicited by body

sway (black arrow) in the sagittal plane, when subjects

have their head directed forwards (Fig. 1A) and in the

frontal plane, when the head is rotated by 90� towards a

shoulder, in the same direction as the body sway (see

Fig. 1B). However, the two illustrated conditions require

d.

A B

DC

Fig. 1. Neck tuning of VS reflexes. (A, B) The relative position of the head with respect to the body changes the coupling between the vestibular

input elicited by body sway (black arrows) and the postural response (white arrows). Note that in both (A) and (B) the direction of head displacement

(and of the elicited vestibular input) is identical but the postural responses must be different in order to maintain balance. (C, D) Effect of a change in

body-to-head position on the preferred direction of the triceps brachii response to tilts in vertical planes. In (C), when the head-and body longitudinal

axes are aligned, the maximal EMG response is obtained for a roll tilt in the frontal plane (indicated by the thick, black arrow), around the longitudinal

head–body axis (y). Tilt around an obliqual axis (y0) elicit a smaller EMG response. In (D) the body was rotated with respect to the head, so that its

axis was parallel to y0. In this position, the maximal response was obtained for a tilt around the y0 axis, while the response to tilt around the y axis getssmaller, so that the preferred response direction (thick, black arrow) remained perpendicular to the longitudinal body axis. The insets represent the

modulations of EMG activity observed during tilt around y and y0.

M. Barresi et al. / Neuroscience 224 (2012) 48–62 49

different postural reflexes (white arrows) in order to

maintain balance. This non-linear interaction of

vestibular and neck signals, which is expected when the

head and neck rotation are not coplanar (Mergner et al.,

1997), can be appreciated when galvanic vestibular

stimulation is delivered with the head oriented in

different positions with respect to the body (Lund and

Broberg, 1983; Britton et al., 1993; Fitzpatrick et al.,

1994). In this instance the central nervous system

interprets the stimulus as a sway in the direction of the

stimulated labyrinth and generates a postural response

in the opposite direction. Both the perceived and the

elicited body sway rotate by the same angle as the

head over the body. The neurophysiological mechanism

of this transformation has been addressed in the

decerebrate cat, where it has been shown that body-

to-head rotation modifies by the same angle the

preferred response direction of the forelimb extensor

muscles response to animal tilt (Manzoni et al., 1998).

In this way the pattern of VS reflexes changes by

changing the position of the head with respect to the

body, so that a given muscle results maximally activated

for a particular direction of the body rather than head

displacement (see Fig. 1C, D). This obviously makes

the reflex appropriate for maintaining body stability. It

has been documented in decerebrate cat (Manzoni

et al., 1998), as well as humans (Kammermeier et al.,

2009) that the tuning exerted on VS reflexes by neck

rotation is abolished by inactivation (Manzoni et al.,

1998) and pathology (Kammermeier et al., 2009) of the

cerebellar vermis, respectively: this means that without

the cerebellum there is a loss of the neck–vestibular

integration allowing to transform information about head

motion into a postural response appropriate to

counteract body sway.

It is of interest that, in decerebrate cat body-to-head

displacement modifies the labyrinthine responses of the

P-cells of cerebellar anterior vermis, similar to what was

observed in the spinal motoneuron: on the average the

preferred response direction of P-cells rotates in the

same direction and by the same amplitude as the body

with respect to the head (Manzoni et al., 1999). Thus,

given the lack of neck tuning on motoneuronal

responses to labyrinthine input which occurs following

cerebellar inactivation, it was proposed that directional

modifications of P-cells may induce similar changes in

the vestibular responses of their target neurons located

within the fastigial and the vestibular nuclei and, as a

consequence, in spinal motoneurons (Manzoni et al.,

1999). Indeed, following changes in the body-to-head

50 M. Barresi et al. / Neuroscience 224 (2012) 48–62

position, directional modifications of responses to

vestibular stimulation have been observed at the level of

the fastigial nucleus (Kleine et al., 2004; Shaikh et al.,

2004), but not within the rostral part of the medial and

lateral vestibular nuclei. Thus, it could be hypothesized

that cerebellar influences on spinal motoneurons related

to neck rotation are exerted through the pontomedullary

reticular formation (RF) (Pompeiano, 1967). This

prompts investigation of the possible role of RF in the

neural processes responsible for changing the effects of

the labyrinthine input from a head-centred into a body-

centred reference frame.

With this aim, we recorded the activity of RF neurons in

urethane anaesthetized rats during labyrinthine stimulation

elicited by animal wobble. These stimuli were run either with

the body and head aligned along the longitudinal axis, as

well as with the body kept at a fixed angle of rotation with

respect to the head in the horizontal plane, towards and

away from the recording side. Since both corticocerebellar

neurons (Manzoni et al., 1999; Kleine et al., 2004; Shaikh et

al., 2004) and spinal motoneurons (Manzoni et al., 1998)

showed directional changes in response to vestibular input

when the head was maintained at a fixed angle with respect

to the body, we expected a similar result at the level of the

RF, which is an integration structure intermediate between

the cerebellum and spinal cord. Since this did not occur our

findings suggest a reconsideration of the role of the

cerebellum in this sensorimotor transformation. A

preliminary report of the present study has been presented

(Barresi et al., 2010).

EXPERIMENTAL PROCEDURES

Animal preparation and unit recording

Experiments were performed in adult, urethane anaesthetized,

Wistar rats (initial dose: 0.75–1.3 g/kg, i.p.). In most of the

experiments a dose of chloralose (60 mg/kg, i.p.) was added

and the dose of urethane was lowered to 0.75 g/kg. All

procedures complied with the National Institute of Health

Guidelines for the Care and Use of Laboratory Animals, as well

as with the European Communities’ guidelines for studies in

animal models (Council Directive of 24/11/1986). In order to

avoid pain and discomfort, the skin and the subcutaneous

tissues incised during the surgery were infiltrated with

novocaine; in addition, the levels of leg withdrawal and corneal

reflex and the electrocardiogram (recorded by needle

electrodes) were monitored. The instantaneous heart rate was

evaluated by analysing the electrocardiogram by a window

discriminator and a rate metre. When corneal and leg

withdrawal reflexes showed any tendency to recover (with

increasing heart rate), additional doses of the anaesthetic were

administered.

Once leg withdrawal and corneal reflexes were depressed by

anaesthesia, the spinous process of the T12–L1 spinal segments

was exposed. The animal’s head (pitched 20� nose-down) was

fixed to a stereotaxic apparatus (David Kopf). The animal’s

body was secured to a spinal cord frame (by a clamp placed

on the T12–L1 spinous process), which was manually rotated

on a horizontal plane around an axis passing approximately

through the first cervical joint and blocked at the new position.

A rubber-heating pad prevented body displacement. The

animals were heated through the body rubber pad, servo-

controlled by a feedback system, in order to maintain the rectal

temperature between 37 and 38 �C. Finally, holes were drilled

in the occipital bone and theta barrel glass microelectrodes

were lowered through the cerebellum. In most caudal

recordings, the microelectrodes were directly inserted into the

medulla’s surface. The recording barrel was filled with a 4 M

solution of NaCl, while the marking barrel contained Pontamine

Sky Blue (5%). The micromanipulator was inclined by 10–25�with respect to the vertical, in order to avoid possible damage

to the transverse sinus. Trackings were performed at distances

from the midline ranging between 0.1 and 2.4 mm. Most of the

penetrations were performed on the left side. On the average,

we recorded about 8 units from each experiment. Marking of

the recording side was performed by iontophoretic application

of the Pontamine Dye (cathodal current, 30 lA, 15–30 min).

A gel of Agar–Agar (2%) covered the exposed brain surface

preventing drying as well as displacement of the cerebellum

and medulla during tilt.

Wobble stimuli and data analysis

The tilting table could rotate around three axes (transverse-pitch,

longitudinal-roll and vertical-yaw) passing through the centre of

the animal’s head (Pompeiano et al., 1997). The roll and pitch

axes of the table were driven with sinusoids (0.156 Hz, 5�) outof phase by 90�. In this way, the animal was submitted to a

‘‘wobble’’ stimulus, i.e. to a tilt of constant amplitude (5�),whose direction rotated at constant velocity (56.2�/s) over the

horizontal plane, either clockwise (CW) or counter clockwise

(CCW) (Schor et al., 1984). The reason for using wobble

stimuli instead of tilt oriented in a specific direction is that the

former approach allows one to define the unit’s response

characteristics with only two stimulation sequences running CW

and CCW. The tilt procedure requires a larger number of

stimulation sequences, whose gain data have to be fitted by a

sinusoidal model. As a consequence, any spontaneous

variability in the gain of tilt response will deeply modify the

estimate of the response direction while this is not the case

when wobble stimuli are utilized. The frequency of stimulation

used allowed for the activation of both otolith and ampullar

afferents, as occurs during natural head movement. During CW

stimuli, units recorded on the right side of the brain were

successively analysed while the animal was tilted side-down

(SD), nose-up (NU), side-up (SU) and nose-down (ND)

(Fig. 2A, upper trace). During CCW stimuli the successive

positions were SD, ND, SD and NU (Fig. 2A). These

sequences reversed when the units were recorded on the left

side. For this reason, in the text, independently of the actual

direction of rotation, the terms CW and CCW rotations will refer

to wobble stimuli eliciting the sequences SD–NU–SU–ND and

SD–ND–SU–NU, respectively. The cell’s firing frequency was

fed to a rate metre and converted to standard pulses by a

window discriminator. The output of the window discriminator

was analysed by a signal processor that counted the number of

spikes generated during 512 sequential intervals (bins) of

100 ms that covered a stimulation sequence of four complete

cycles of wobble in a given (CW or CCW) direction (128 bins

for each cycle). The period of the individual wobble cycle was

6.4 s. The instrument averaged the responses to four

stimulation sequences and generated four-cycle sequential

pulse density histograms (SPDH). For illustration purposes, the

first and the second cycles of the SPDHs were averaged with

the third and fourth, respectively, to obtain a two-cycle display

(see Fig. 2B). The SPDHs were submitted to a Fast Fourier

analysis providing the base frequency (BF, in impulses/s), i.e.

the mean discharge rate of the unit during wobble, the gain (in

imp/s/�, indicated as GCW and GCCW, for CW and CCW

rotation, respectively) and the phase angle (in arc degrees) of

the first harmonic component (0.156 Hz) of the averaged

responses. The phase angle corresponded to the direction of

head tilt giving rise to the peak discharge of the unit during

wobble: 0� and 180� indicated SD and SU displacement, while

A B C

Fig. 2. Response of a representative reticular neuron to animal wobble in CW and CCW direction. (A) Upper plot. The direction of the animal tilt was

evaluated with respect to a reference system where 0� and 180� represented tilts along the frontal plane, towards and away from the recording side,

while 90� and �90� represented downwards and upwards displacements along the sagittal plane. The sequential directions of the animal tilt

occurring during wobble in CW and CCW directions can be obtained by following the corresponding black arrows. Lower plot. The directions of the

CW (DCW) and CCW (DCCW) responses have been reported as solid arrows. The dashed arrow represents the direction of the response vector

evaluated for the same neuron (Hmax). Its direction (�99.7�) corresponds to the midposition between the CW and CCW discharge maxima. SD:

side-down; NU: nose-up; SU: side-up; ND: nose-down. (B) Sequential pulse density histograms (SPDHs) showing the unit activity recorded from a

reticular neuron during two sequential periods of wobble either in CW (upper trace) or CCW (lower trace) direction. Each trace is the average of eight

superimposed sweeps. The response gain corresponded to 0.878 imp/s/� and to 1.027 imp/s/� for the CW and the CCW response, respectively,

while the response direction (i.e. the direction of tilt giving rise to the maximal activity of the cell) was �113.3� for the CW and �86.0� for the CCW

response. These values are indicated by the vertical, dashed lines. The sinusoids-fitting unit activity traces correspond to the first harmonic of the

response, as evaluated by Fourier analysis. (C) Illustration of experimental animal posture paradigms. Upper figure: control position with the head

and the body axes aligned. Middle and lower figures: body-rotated positions, with the trunk axis displaced to the right and left side, respectively.

M. Barresi et al. / Neuroscience 224 (2012) 48–62 51

90� and �90� indicated ND and NU displacements, respectively

(see Fig. 2A). SD and SU positions were referred to the side of

unit recording. In the text the phase angle will be indicated as

‘‘response direction’’ (DCW and DCCW) or as ‘‘position of

maximal discharge’’.

In order to assess the unit responsiveness, the coherence

and the signal-to-noise ratio were computed. The coherence

corresponded to the squared amplitude of the ratio between the

vectorial and the algebraic mean of the response amplitudes (in

impulses/s) relative to the four wobble cycles included in the

stimulation sequence. The value 1 was obtained only when

these two gain and phase values were equal. The signal-

to-noise ratio corresponded to the ratio between the amplitude

of the fundamental harmonic of the response and the root-

mean-square amplitude of all harmonics above the second. In

general, criteria for considering a record as responsive were

signal-to-noise ratio and coherence > 0.5.

Experimental protocol

For each unit, responses to wobble in both CW and CCW

direction were recorded in two different conditions: (1) control,

with trunk and head longitudinal axes aligned to each other

(see Fig. 2C, upper panel) and (2) trunk rotated, with a static

rotation of the trunk with respect to the head (2.5–30�) in either

direction of the horizontal plane (see Fig. 2C, middle and lower

panels). The rotation was manually imposed to the spinal frame

and measured by an arc goniometre over which the frame was

sliding. When possible, the response in the control and/or in

the trunk-rotated condition was tested more than once.

Response vectors

The direction of the response to CW and CCW wobble stimuli

depends on the tilt direction giving rise to the best response of

the neuron (spatial property) as well as on the temporal relation

between response and stimulus peaks (temporal property).

Spatial and temporal response properties of individual neurons

can be disentangled only by taking into account the responses

to both CW and CCW stimuli (Schor et al., 1984). This analysis

allows one to evaluate response vectors, which are defined by

gain (G), direction (spatial property) and phase (temporal

property) components (Schor and Angelaki, 1992; Bush et al.,

1993).

The possibility of defining response vectors is related to the

characteristics of vestibular receptor responses. It is known, in

fact, that otolith receptors show a response gain to tilt stimuli in

vertical planes, which is proportional to the cosine of the angle

between the direction of tilt (i.e. the axis perpendicular to the

plane of rotation) and the axis joining the bundle of stereocilium

to the kinocilium (polarization axis). So, the response gains to

tilt stimuli can be obtained by projecting, along the tilt direction

52 M. Barresi et al. / Neuroscience 224 (2012) 48–62

(a), a response vector having a length corresponding to the

direction of the polarization axis (h), of the gain obtained for

stimuli oriented along the axis (G) (Schor and Angelaki, 1992).

When sinusoidal tilt stimuli are utilized the relation between

neural response (R), direction of tilt (a) and polarization axis (h)corresponds to:

R ¼ G � cosðh� aÞ � senðxtþuÞ;x ¼ 2pf ðf ¼ frequency of tiltÞ;u ¼ phase of receptor’s response:

Since the response of ampullar receptors is proportional to

the cosine of the angle between the plane of tilt rotation and

that of the canal, their response can be described in a similar

way.

It can be shown that, when signals from receptors having the

same preferred direction and/or the same phase converge at a

single unit level, the neuronal responses to tilt stimuli can be

predicted by a single response vector (maximal sensitivity vector,

Smax): in this instance the responses to CW and CCW wobble are

of about the same amplitude (Schor and Angelaki, 1992). The

direction (orientation) of Smax (hmax) corresponds to the direction of

stimulus giving rise to the maximal response (preferred direction),

while its gain (Gmax) and temporal phase (umax) are the gain and

phase observed for a stimulus in the preferred direction. In the

present experiments, the temporal phase was evaluated with

respect to the animal position, negative and positive values of umax

representing lags and leads of the peak of the unit activity with

respect to the peak of the animal displacement, respectively. Gmax,

hmax and umax are evaluated as follows1: Gmax = (GCW +GCCW)/2,

hmax = (DCW + DCCW)/2, umax = (DCW� DCCW)/2 (Schor et al.,

1984; Schor and Angelaki, 1992). As stated above the modulation in

firing rate during a sinusoidal tilt in a given direction can be predicted

by the projection of the cell response vectors along the stimulus

direction. Obviously, the response is maximal for tilt in the direction of

hmax and null for the perpendicular direction: these units are indicated

as bi-directional, ‘‘narrowly tuned’’ units. On the other hand, when

signals from receptors endowed both with a different preferred

direction and a different phase converge at single unit level

(spatiotemporal convergence, STC), the unit responses to tilt are

better predicted if a second vector Smin, orthogonal to Smax is

included in the model. In this instance, the response gains to wobble

in CW and CCW direction are different. These bi-directional units are

indicated as ‘‘broadly tuned’’ neurons, since they show a minimal but

significant response to tilt in the direction perpendicular to hmax (Schor

and Angelaki, 1992; Bush et al., 1993). The phase of the response

along this direction (umin) leads umax by 90�. The ratio between the

gain of Smin and Smax is called tuning ratio and is taken as a measure

of the degree of STC occurring at single unit level. Units unaffected

by wobble in a given direction (unidirectional units) represent the

extreme case of ‘‘broadly tuned’’ units: they are expected to show the

same response gain for all the directions of tilt. In these instances,

Smax is equal to Smin and its gain corresponds to half of the response

gain obtained for the direction of wobble (either CW or CCW) able to

elicit the cell response (Angelaki, 1992). For the latter units, the

phase of the response changes linearly with the direction of tilt

(Angelaki, 1991).

Statistical analysis

Statistical comparison between control and post-rotation data

was performed through repeated measures analysis of variance

(ANOVA). The significance level was set at P< 0.05.

1 All the following computations are relative to units recorded on theright side. For left side recorded units CW and CCW parameters haveto be reciprocally exchanged.

Histology

The localization of recorded units was evaluated on the basis of

the position of Pontamine Blue spots, identified on sagittal

sections of brainstem stained with neutral red (see Fig. 3).

Marking points allowed us to establish the position of selected

points as well as the orientation of the tracks on the section.

The relative position of each recorded neuron was evaluated by

comparing the corresponding stereotaxic coordinates with

those of the marking spots.

RESULTS

Recorded population

The activity of 139 neurons was recorded from the RF during

wobble stimuli, with the animal in the control position; in

particular, 130 neurons were recorded during CW and 97

during CCW rotations. All neurons were located in a region

ranging from the midline to lateral 2.4 mm, between P levels

�0.4 mm and �5.6 mm from interaural, that encompassed

the paramedian reticular nucleus, the gigantocellular

reticular nucleus and the parvocellular reticular nucleus. On

the whole, 33% of the recorded neurons were affected by

CW and/or CCW rotations. The histological localization of

both responsive and unresponsive units is shown in Fig. 3A.

Average values for base frequency, gain, coherence and

SN ratio are given in Table 1 for both CW and CCW

responses, together with the distribution of the

corresponding response directions, that could be related to

SD (�45� < D<45�), SU (D>135� or <�135�), ND

(45� < D<135�) and NU (�135� < D<�45�).Response gains and direction were not correlated.

Moreover, gains were not related to the corresponding base

frequency values.

The characteristics of Smax response vector could be

determined only in the limited number of units tested

during CW and CCW rotation and responsive to both

stimuli (n= 11). In particular, 6 units showed responses

of comparable amplitude, being therefore classified as

narrowly tuned neurons, while 5 units characterized by

gain ratios of the CW to the CCW response larger than

1.22 or smaller than 0.82 were considered broadly

tuned neurons (Angelaki, 1992). Ten additional units

could be selectively affected by CW (N= 4) or CCW

wobble (N= 6), being, therefore, unidirectional units.

The difference in responsiveness to CW and CCW

rotations could be verified by repeated tests in 4 of

these units. Average gain values for these populations

are given in Table 2, together with the distribution of

Smax (bidirectional units) or CW/CCW response

(unidirectional units) direction. Bidirectional units had

significantly larger gains than unidirectional cells.

Effects of ipsilateral trunk rotation on responsecharacteristics of RF units to CW and CCW wobble

Thirty-one units that responded in the control position

were also tested following ipsilateral trunk rotation.

Twenty-one of them were analysed only during CW

rotation, 10 during CCW rotation only and only two for

both directions of rotation. The histological localization

of these neurons is shown in Fig. 3B. Trunk

A B

Fig. 3. Histological localization of the recorded neurons. The histological localization of the recorded neurons has been reported on two histological

sections corresponding to a laterality of 0.40 and 1.90 mm, respectively. From left to right: rostro-caudal; from top to bottom: dorsal–ventral. (A)

Units responsive (d) and unresponsive (�) to labyrinthine stimulation are shown. (B) Histological localization of the units affected by labyrinthine

stimulation whose response to wobble has been tested following the ipsilateral trunk rotation. Black and white symbols represent CW and CCW

recordings, respectively. Downward- and upward-oriented triangles refer to the units whose response gain was decreased and enhanced by

ipsilateral body rotation, respectively. Open and closed circles represent units whose gain was unaffected by body rotation. I: lobule one of Larsell;

IX: lobule nine of Larsell; X: lobule ten of Larsell; 6: abducens nucleus; 7: facial nucleus; 7n: facial nerve; 10: dorsal motor nucleus of the vagus; 12:

hypoglossal nucleus; Amb: nucleus ambiguous; Cop: copula of the piramis; DMSp5: dorsomedial spinal trigeminal nucleus; DPGi: dorsal

paragigantocellular nucleus; ECu: external cuneate nucleus; Gi: gigantocellular reticular nucleus; Gia: gigantocellular reticular nucleus, pars alpha;

GiV: gigantocellular reticular nucleus, pars ventralis; Gr: gracile nucleus; IO: inferior olive; LPGi: lateral paragigantocellular nucleus; LRt: lateral

reticular nucleus; LVe: lateral vestibular nucleus; MdD: dorsal medullary reticular field; Me5: mesencephalic trigeminal nucleus; MVe: medial

vestibular nucleus; nx: nucleus X; PCRt: parvocellular reticular nucleus, PMn: paramedian reticular nucleus, PnC: pontine reticular nucleus, ventral

division; PnV: pontine reticular nucleus, ventral division; Prh: prepositus hypoglossi; Py: pyramidal tract; Ro: nucleus of Roller; scp: superior

cerebellar peduncle; SGe: supragenual nucleus; Sol: solitary tract nucleus; Sp5C: spinal trigeminal nucleus, caudal part; SpVe: spinal vestibular

nucleus; SuVe: superior vestibular nucleus; tz: trapezoid body.

M. Barresi et al. / Neuroscience 224 (2012) 48–62 53

displacement significantly modified the characteristics of

the response to wobble. Changes were observed in

base frequency, response gain and, less frequently,

also in the response direction. The reliability of these

changes was assessed by repeating the tests under

control and trunk-rotated positions. In addition, we

compared the changes in response properties induced

in all the tested units by trunk rotation (at the largest

angle of body rotation tested) with those observed when

the wobble was repeated twice at the same body

position (usually with the head and the trunk colinear).

These changes were expressed as the absolute value

of the ratio (control � trunk rotated)/[(control + trunk

rotated)/2]. Obviously, those units that were silenced by

body rotation were considered as having BF and gains

zero in the body-rotated position. As shown in Table 3A,

the changes in BF, gain, SN ratio and coherence

induced by trunk rotation in the whole population

analysed, were significantly larger than those expected

by chance.

Data relative to the units that effectively changed their

BF are shown in Table 4. BF changes occurred in about

half of the units during CW and CCW stimuli. During

CW rotation, only activity drops were observed, while,

during CCW rotation, drops and enhancements had the

same chance to occur. It is interesting that three

‘‘dropping’’ units were silenced by body displacement.

Moreover, when trunk rotation was raised from

Table 1. Average (±standard deviation) has been evaluated for the different parameters of labyrinthine responses obtained during CW and CCW

wobble stimuli. SD: �45� < response direction < 45�; ND: 45� < response direction < 135�; SU: response direction > 135� or < �135�; NU:

�135� < response direction < �45�

Base frequency unresponsive units Base frequency responsive units Gain Coherence SN ratio

CW 14.6 ± 13.0 17.2 ± 10.7 0.601 ± 0.543 0.81 ± 0.12 100.9 ± 69.1

N= 96 N= 34 N= 34 N= 34 N= 34

CCW 16.3 ± 15.1 17.4 ± 10.1 0.724 ± 0.920 0.75 ± 0.16 100.5 ± 92.8

N= 74 N= 23 N= 23 N= 33 N= 33

Distribution of response direction

SU NU SD ND

CW N= 8 N= 5 N= 6 N= 15

CCW N= 8 N= 8 N= 3 N= 4

Table 2. Values represent the average ± standard deviation of the gain of Smax vector (Gmax). CW unidir and CCW unidir are the units affected only

during CW and CCW rotations, respectively. Numbers refer to the direction of the response vector (bidirectional units) or to the direction of the CW/

CCW responses (unidirectional units). The asterisks indicated a significant difference between bidirectional and unidirectional cells. See text for further

explanations

Gain of Smax Direction of Smax

SD SU ND NU

Narrowly tuned 0.69 ± 0.62 n= 2 n= 2 n= 1 n= 1

n= 6

Broadly tuned 0.78 ± 0.79 n= 1 n= 3 n= 1 n= 0

n= 5

All bidirectional 0.73 ± 0.67 n= 3 n= 5 n= 2 n= 1

n= 11

Direction of CW/CCW response

CW unidir 0.18 ± 0.07 n= 1 n= 1 n= 2 n= 0

n= 4

CCW unidir 0.29 ± 0.25 n= 1 n= 1 n= 3 n= 0

n= 5

All unidir 0.22 ± 0.20 ⁄P< 0.004 n= 2 n= 2 n= 5 n= 0

n= 9

54 M. Barresi et al. / Neuroscience 224 (2012) 48–62

0� (control) to 10� and 20�, the BF was highly correlated

with the latter parameter both in the units showing drops

(n= 4) or enhancements (n= 2), as indicated by

regression of all the experimental points. In this

analysis, individual unit’s data were expressed as a

percentage with respect to the average of the three

values obtained at 0�, 10� and 20�. The regression lines

corresponded to Y= �2.9X+ 128.5 (r= 0.88,

P< 0.0005) for ‘‘dropping’’ units and to Y = 1.8X+

81.5 (r= 0.98, P< 0.0005) for units with BF

enhancements.

As shown in Table 4A, gain changes were more

frequent than base frequency changes: yet, drops

predominate over enhancements during CW rotation,

while they were as numerous as the latter during CCW

rotation. On the whole, 9 units became unresponsive to

wobble following trunk rotation. Most of these neurons

could be tested once more in the control condition, thus

indicating that loss in responsiveness was not due to

deterioration of unit recording. Representative examples

of neurons showing drops and enhancements in

response gain following body rotation are shown in

Fig. 4A and B, respectively. Similar to BF, gain was

highly correlated to the amplitude of body rotation both in

units with drops (n= 7) or enhancements (n= 5). The

regression lines corresponded to Y= �4.0X+ 139.7

(r= 0.80, P< 0.0005) for ‘‘dropping’’ units and to

Y = 4.3X+ 56.8 (r= 0.86, P< 0.0005) for units with

gain enhancements.

Although changes in gain and base frequency were

significantly correlated (r= 0.687, P< 0.0005), only

47.2% of the gain variability could be attributed to the

corresponding variation in base frequency. In several

units, gain changes were not coupled with base

frequency modifications, or the two changes were of

opposite signs. On the other hand a stronger correlation

could be observed between changes in gain and

modifications in SN ratio and coherence (SN ratio:

r= 0.98, P< 0.0005; coherence: r= 0.93,

P< 0.0005), as could be expected by the fact that all

these parameters reflect the unit’s responsiveness.

As to the directional changes induced by ipsilateral

trunk rotation (i.e. the difference between trunk-rotated

and control conditions), their absolute value was not

significantly different from that attributable to chance

(Table 3A); however, in at least 5 units the changes in

response direction induced by trunk rotation could be

validated by repeated testing of the cell. No significant

Table 3. Average ± standard deviation values of the changes (absolute values) in response parameters obtained by repeating the same test twice and

by rotating the trunk with respect to the head, towards and away from the recording side. Asterisk represents significant differences with respect to the

simple test repetition

Changes induced by test repetition (A) Changes induced by ipsilateral trunk

rotation

(B) Changes induced by ipsilateral trunk

rotation

n= 43 n= 33 n= 24

Base frequency 20.7 ± 20.7%⁄ 44.0 ± 55.5% ⁄P< 0.013 25.5 ± 32.7% NS

Gain 23.1 ± 17.9%⁄ 79.8 ± 60.1% ⁄P< 0.0005 77.0 ± 53.8% ⁄P< 0.0005

SN ratio 30.1 ± 26.4%⁄ 79.4 ± 55.8% ⁄P< 0.0005 80.0 ± 51.8% ⁄P< 0.0005

Coherence 14.8 ± 14.2%⁄ 66.5 ± 69.7% ⁄P< 0.0005 37.3 ± 47.6% ⁄P< 0.005

n= 43 n= 24 n= 15

Direction 19.9 ± 21.0� 26.1 ± 34.0� NS 72.8 ± 49.1� ⁄P< 0.0005

M. Barresi et al. / Neuroscience 224 (2012) 48–62 55

differences in the directional changes induced by

ipsilateral body rotation were found between CW and

CCW responses. Units showing directional modifications

showed either a progressive rotation of their response

direction by increasing the angle of body rotation, or a

saturation effect.

The effects of ipsilateral trunk rotation could be tested

in 9 units whose discharge was not modulated by wobble

stimuli in the control position (signal-to-noise ratio and

coherence < 0.5). While in 6 units body rotation left the

unit behaviour unmodified, in 3 units a clear modulation

with both coherence and signal-to-noise ratio > 0.5

showed up following ipsilateral body rotation: one of

these, in particular, was completely silent in the control

condition.

It is worth of note that units showing BF/gain drops

and enhancements could be found along the whole

extent of the explored medullary region, while neurons

with directional changes were medially located, four out

of five being within the paramedian reticular nucleus.

Effects of the contralateral trunk rotation on responsecharacteristics of RF units to CW and CCW wobble

On the whole, 20 units affected by wobble stimuli in the

control condition could be tested during contralateral

trunk rotation either in the CW (n= 14), in the CCW

(n= 10) or in both directions (n= 4).

Following trunk rotation towards the side opposite to

that of unit recording, individual units could show

changes in BF as well as in gain and direction of their

responses to wobble stimuli. As shown in Table 3B,

when the whole population of tested units was

considered, the changes induced by trunk rotation were

significantly larger than those expected by chance for all

the parameters, with the exception of BF. Data relative

to neurons showing reliable changes in BF and gain

following trunk rotation, together with the corresponding

average values are given in Table 4B.

When the amplitude of body rotation was increased

from 10� to 20�, units with BF enhancements showed a

saturation of the effect: only one ‘‘dropping’’ unit could

be tested in the same way, which showed a progressive

increase in BF by increasing the amplitude of body

displacement.

As observed during ipsilateral rotation, changes in

responsiveness were more frequent than those in BF

(Table 4B); however, they were independent from BF

changes. Gain co-varied with the amplitude of trunk rotation

both in ‘‘dropping’’ units (n=3), as well as in units with

enhancements (n=3). The regression lines corresponded

to Y=�8.1X+181 (r=0.86, P< 0.003) for ‘‘dropping’’

units and to Y=4.2X+58.3 (r= 0.96, P<0.0005) for

units with gain enhancements. The changes in coherence

and SN ratio were strongly correlated with the

corresponding gain modifications (coherence: R=0.65,

P<0.004; SN: R=0.94, P<0.005). It is of interest that,

following contralateral trunk displacement, 8 units became

unresponsive to CW (n=5) or CCW (n=4) rotation. Five

of those units could be tested once more in the control

condition, thus indicating that loss in responsiveness was

not due to deterioration of unit recording.

Finally, directional changes could be observed in 9 of

the tested neurons: one of these units has been shown in

Fig. 5. As shown in Fig. 6A, directional changes appeared

to be of rather large amplitude with respect to the angle of

trunk rotation (5–20�). In this respect, no significant

differences could be observed between CW and CCW

recordings. When the angle of body rotation was

increased from 10� to 20�, the directional change could

either increase or be left unmodified (Fig. 6B).

Four units that were not affected by wobble stimuli

were also tested following trunk rotation: one of these

neurons became responsive at the new position.

Units showing BF and gain changes during

contralateral trunk rotation could be recorded all over the

explored reticular region, while those showing directional

modifications could be found exclusively within the

paramedian reticular nucleus.

Single-unit behaviour during body rotation inopposite directions

A limited number of units were tested during wobble stimuli

in the control position and following body rotation in

opposite directions. On the basis of gain changes

neurons were subdivided into different groups. In a first

group of eight cells, mostly unresponsive at �20� (and

sometimes at �10�) of (contralateral) body rotation, an

increase in gain was observed by rotating the body

towards the recording site (positive gain trend). These

units have been shown in Fig. 7A, where the gain has

been plotted as a percentage of the average

values relative to all the trunk positions tested. Gain

Table 4. Changes in response Gain and BF elicited by ipsilateral (A) and contralateral (B) body displacement. The table reports the numbers of neurons

showing different patterns of changes in base frequency and/or gain induced by trunk rotation in all the CW and CCW responses

Unaffected Drops Increases

(A) Ipsilateral trunk rotation

Basal frequency CW n= 12 n= 9

�106.8 ± 72.4

n= 0

CCW n= 5 n= 3

�48.0 ± 13.6

n= 4

42.4 ± 7.7

All n= 17 n= 12

�92.1 ± 67

n= 4

42.4 ± 7.7

Gain CW n= 3 n= 13

�117.8 ± 57.5

n= 5

42.1 ± 20.9

CCW n= 3 n= 4

�102.1 ± 39.7

n= 5

79.7 ± 51.7

All n= 6 n= 17

�114.1 ± 53.1

n= 10

60.9 ± 42.1

(B) Contralateral trunk rotation

Basal frequency CW n= 7 n= 4

�52.0 ± 52.1

n= 3

23.6 ± 6.7

CCW n= 8 n= 1

�124.3n= 1

34.6

All n= 15 n= 5

�66.1 ± 55.6

n= 4

26.4 ± 7.8

Gain CW n= 4 n= 7

�106.7 ± 63.0

n= 3

79.4 ± 29.5

CCW n= 2 n= 7

�99.8 ± 36.1

n= 1

70.1

All n= 6 n= 14

�103.2 ± 49.4

n= 4

77.1 ± 24.5

56 M. Barresi et al. / Neuroscience 224 (2012) 48–62

modifications were highly correlated to body rotation

(r= 0.87, P< 0.0005, Y= 4.4X+ 100). Changes in

responsiveness were not attributable to parallel changes

in unit firing rate, since no correlation existed between

the two parameters: within this population, in fact, BF

could increase (n= 5), decrease (n= 3) or stay

unmodified when the trunk was rotated from contralateral

to ipsilateral side. Moving from contralateral to ipsilateral

body displacements, the position of CW (n= 3) and

CCW (n= 5) the discharge maxima of these units

showed the tendency to rotate in the same direction as

the body (Fig. 7B): on the whole, a significant correlation

was found between the changes in response direction

and the angle of body rotation (r= 0.58, P< 0.003,

Y= 3.17X � 2.64). It is of interest that this correlation

persisted when the unit characterized by the largest

directional variation (about 155� at 10� and 20� of body

rotation) was excluded from the analysis (r= 0.528,

P< 0.014, Y= 1.604X � 3.50). As expected, directional

variation was significantly correlated to the gain change

(r= 0.547, P< 0.006, Y= 0.553X � 54.9). Six of these

neurons were located in the paramedian reticular

nucleus while two of them in the medial portion of the

gigantocellular field.

In a second group of three cells shown in Fig. 7C, a

progressive decrease in the gain of the CW response

from a �20�- to a 20�-body position was observed

(negative gain trend). None of these units could be

tested during CCW rotation. Gain changes were

strongly correlated to body displacement (r= 0.82,

P< 0.0005, Y= �2.4X+ 101.7). The changes in gain

were not attributable to changes in neuronal excitability,

since, in these units, body rotation did not induce any

change in BF. As shown in Fig. 7D, following

contralateral body rotation, these neurons showed

displacement of their discharge maxima towards the

recording side, while ipsilateral rotation induced minor

directional changes in the same direction, thus giving

rise to a significant correlation between changes in

response direction and body rotation (r= 0.68,

P< 0.016, Y= �1.79X+ 33.1). All these units were

located in the paramedian reticular nucleus. Finally, in

five ‘‘tuned’’ neurons (all of them located within the

paramedian reticular nucleus), the gain decreased

following the bilateral trunk rotation (Fig. 7E). In

particular in four ‘‘silenced’’ neurons, all tested during

CW rotation (Fig. 7E), small body rotations in a given

direction abolished completely the spontaneous activity

and the responsiveness to tilt. For all these four

neurons, body rotation in the direction opposite to that

silencing the cell did not greatly affect the basal activity

but greatly depressed the unit responsiveness that

could be even lost. When the units preserved their

responsiveness following body rotation, the position of

CW discharge maxima always rotated in the same

direction as the body (Fig. 7F).

The directional modifications induced by trunk rotation

in the three groups of neurons were compared by a 3 (unit

type) � 2 (direction of body rotation) experimental design

(ANOVA), which indicated a significant type � direction

A B

Fig. 4. Effect of ipsilateral trunk rotation on the responses of two representative reticular units to wobble stimuli. SPDHs obtained during animal

wobble (two successive periods) in 2 units (A, B) for different positions of the head with respect to the trunk. Each trace is the average of eight sweeps.

Units were recorded on the left side. (A) Responses obtained during CW rotation. Top trace: in the control position the response gain was to 0.75 imp/

s/�, the direction �62.4�, while SN ratio and coherence corresponded to 3.50 and 0.93, respectively. Middle trace: 10� of trunk rotation towards the

recording side (left) decreased the gain to 0.34 imp/s/�, while the direction was virtually unmodified (�64.3�). SN ratio and coherence were 1.25 and

0.97, respectively. Bottom trace: back in the control position the gain raised again to 0.58 imp/s/�, which was close to the initial value (top trace).

Direction, SN ratio and coherence values were�70.6�, 1.81 and 0.94, respectively. (B) Responses obtained during CCW rotation. Top trace: at 20� oftrunk rotation towards the recording side (left) the response gain of the unit was to 5.170 imp/s/�, while its direction value corresponded to�41.9�. SNratio and coherence were 3.85 and 0.95, respectively. Second trace from above: when trunk and head longitudinal axes were aligned in the control

position, the gain decreased to 1.363 imp/s/�. Direction, SN ratio and coherence were�80.1�, 1.28 and 0.72, respectively. Third trace from above: at

10� of trunk rotation towards the recording side the gain raised to 4.243 imp/s/�. Direction, SN ratio and coherence were �39.2�, 3.01 and 0.96,

respectively. Bottom trace: back in the 20� position the gain was 5.936 imp/s/�, which was close to the initial value (top trace). Direction, SN ratio and

coherence values were �43.4�, 3.49 and 0.98, respectively. In both (A) and (B), the sinusoids-fitting unit activity correspond to the first harmonic of

the response, as evaluated by Fourier analysis. ND: nose-down; NU: nose-up; SD: side-down; SU: side-up.

M. Barresi et al. / Neuroscience 224 (2012) 48–62 57

interaction (P< 0.007). Post-hoc comparison indicated a

significant difference of units with negative trend with

respect to both positive trend (P< 0.005) and tuned

units (P< 0.002) during contralateral body rotation.

Finally, one additional unit unaffected by CW wobble

in the control position became responsive following body

rotation in both directions: yet, the gain changes were

independent of the corresponding base frequency

changes.

Changes in response vectors elicited by trunkrotation

Only 7 units could be tested for both CW and CCW

rotation in both the control and the trunk-rotated

condition, which does not allow to draw a definitive

conclusion on the behaviour of the whole population.

However, it was possible to show that trunk rotation

may modify the gain of the response vector in

(5/7 units) and induces a very large change in its

direction (2/7 units).

DISCUSSION

Changes in response properties of RF neuronselicited by trunk rotation

The present findings indicate that the response properties

of some RF neurons to labyrinthine stimulation are

modified by the changes in the relative position of the

head with respect to the body. Thus, the RF may play a

key role in the process that transforms the labyrinthine

volleys related to head motion into postural responses

that stabilize the position of the body in space (Lund and

Broberg, 1983; Britton et al., 1993; Fitzpatrick et al., 1994).

Since wobble at 0.15 Hz may affect canal and otolith

afferents (Kasper et al., 1988; Bolton et al., 1992), this

tuning action could be exerted on both kinds of

information. However, about 64% of RF neurons do not

show convergence of the two inputs (Bolton et al., 1992).

The integration properties of cerebellar neurons

(Manzoni et al., 1999; Kleine et al., 2004; Shaikh et al.,

2004) have been considered the neural substrate of the

change in the reference frame for vestibular input, which

is coded in a head-centred reference system, but must

generate body-centred actions and perceptions

(Mergner et al., 1997). It was assumed (Manzoni et al.,

1999) that the changes occurring in the cerebellum

(Manzoni et al., 1999; Kleine et al., 2004; Shaikh et al.,

2004) could modify the motor output through the RF

(Pompeiano, 1967). So, we were expecting a close

correspondence between the response modifications

observed in cerebellar structures (which concerned

mainly the neuronal response direction) and those

occurring within the RF. However, within the RF, the

most common modifications elicited by trunk rotation

concerned base frequency and gain, while changes in

the direction of labyrinthine responses occurred only in

a minority of units.

A B

Fig. 5. Effects of contralateral trunk rotation on the response direction of a representative reticular unit to animal wobble. (A) SPDH showing the

response of a reticular unit to CCW animal wobble (a single period) for different positions of the head with respect to the trunk. Each trace is the

average of 16 sweeps. In the upper trace the animal had its trunk aligned to the longitudinal head axis. The response gain and direction

corresponded to 0.64 imp/s/� and to �33.9�, respectively. In the middle trace the animal had the trunk rotated contralaterally with respect to the

recording side (left). The response gain (0.31 imp/s/�) was decreased with respect to control, while its direction was shifted by about 100� (�136.3�).In the bottom trace, back to the control position, gain (0.62 imp/s/�) and direction (�58.0�) values were close to those originally recorded in the same

position. (B) Same unit as in (A), but showing responses to CW animal wobble. In the upper trace the animal had its trunk aligned to the longitudinal

head axis. The response gain and direction corresponded to 0.64 imp/s/� and to 54.7�, respectively. In the middle trace the animal had the trunk

rotated contralaterally with respect to the recording side (left). The response gain (0.56 imp/s/�) changed little with respect to control, while its

direction (179.7�) shifted by more than 100�. In the bottom trace, back to the control position, gain (0.86 imp/s/�) and direction (58.9�) values were

close to those originally recorded in the same position. In (A) and (B), the sinusoids-fitting unit activity correspond to the first harmonic of the

response, as evaluated by Fourier analysis. ND: nose-down; NU: nose-up; SD: side-down; SU: side-up.

58 M. Barresi et al. / Neuroscience 224 (2012) 48–62

In several neurons, gain changes did not stem from

modifications in cell’s excitability, being therefore

attributable to specific modifications in the transfer of

labyrinthine information from vestibular receptors to

reticular neurons. Given the difference in neck tuning of

labyrinthine responses observed in the cerebellum and

in the RF, the gain changes of RF neuronal responses

cannot be the mere consequence of corresponding

modifications occurring in the cerebellum.

We may wonder whether directional modifications of

P-cells may be transformed in gain modifications at RF

level, due to the complex recurrent interactions existing

between cerebellar structures, vestibular nuclei and RF.

The lack of neural network modelling studies does not

allow to exclude this hypothesis, although it has been

shown, in analogous sensorimotor transformations (see

the next section) that directional modifications in

response properties may arise from processing of gain

modifications.

There are pathways other than the cerebellar ones that

could be transfer labyrinthine information tuned by neck

rotation to RF neurons. Reticular neurons may receive

labyrinthine information through spinal afferents

(Petrovicky, 1976), similar to what has been shown for the

lateral reticular nucleus (LRN) (Coulter et al., 1976). Since

spinoreticular neurons also receive somatosensory

information (Grant et al., 1966), the labyrinthine signal

could be tuned by neck rotation at this level. Finally,

preliminary results indicate that, within the vestibular nuclei,

the gain of labyrinthine responses is tuned by the neck’s

input (Barresi et al., 2010). Thus, tuning of reticular neuron

responses could arise from the properties of vestibulo-

reticular projection neurons (Ladpli and Brodal, 1968;

Peterson and Abzug, 1975). Since neck tuning of VS

reflexes is abolished by cerebellar inactivation (Manzoni et

al., 1998; Kammermeier et al., 2009), all these hypotheses,

would imply a cerebellar control over the nervous

structures that transfer labyrinthine information tuned by

neck rotation to RF neurons.

The cerebellum and the reference frame for VSreflexes

In summary, the present findings do not support the

hypothesis that changes in the properties of labyrinthine

responses of cerebellar P-cells are simply transferred at

the level of RF (Manzoni et al., 1999). They prompt a

re-consideration of the role of cerebellar and brainstem

structures in the modulation of the vestibular reflexes.

The basic circuitry of the vestibular reflex pathways is

formed by primary sensory neurons that synapse on VS

nuclei projecting to the spinal cord (Uchino et al., 2000).

Labyrinthine information may also impinge on reticular

neurons through the vestibular nuclei (Ladpli and Brodal,

1968; Peterson and Abzug, 1975). The cerebellar vermis

of the anterior lobe receive parallel labyrinthine

A

B

Fig. 6. Effects of contralateral trunk rotation on the direction of CW

and CCW responses. (A) The response directions to CW (d) and

CCW (h) wobble stimuli obtained with the trunk rotated have been

plotted as a function of the corresponding values evaluated in the

control position. The dotted line represents equal values of the two

parameters. 0�, 90�, 180� and �90� represent animal tilts directed

towards the side of unit recording, downwards, towards the opposite

side and upwards. (B) Changes induced in the response direction of 2

units (tested during CW rotation), by increasing the amplitude of

contralateral body rotation.

M. Barresi et al. / Neuroscience 224 (2012) 48–62 59

information (Denoth et al., 1979; Pompeiano et al., 1997;

see Ito, 1984) and controls both VS and reticular neurons

(Pompeiano et al., 1967). At muscle level, the effect of

changing the trunk-to-head position consists of a

modification of the preferred direction of response to

labyrinthine stimulation, which keeps invariant its position

with respect to the trunk (Manzoni et al., 1998). Since the

same pattern is also observed at cerebellar cortical and

nuclear neurons (Manzoni et al., 1999; Kleine et al.,

2004; Shaikh et al., 2004), its apparent lack at the

intermediate station of the RF is somehow puzzling.

There are two possible explanations for this finding. First,

it could be postulated that the recorded reticular neurons

were not projecting to the spinal cord, but to the

cerebellum. In this instance, the occurrence of neurons

with response gain modulated by neck rotation could

make sense. In fact, both theoretical views and

experimental studies (Brotchie et al., 1995; Salinas and

Abbott, 1995; Duhamel et al., 1997) indicate that, in the

changes of reference frame occurring in visuomotor

transformation (Roll et al., 1991), the first step of neural

processing consists of a gain modulation of retinocentric

visual response by the regulatory input (Brotchie et al.,

1995; Duhamel et al., 1997), a phenomenon which is

indicated as ‘‘gain field’’. Only at later stages visual

responses become related to the position of the stimulus

in space and not to the position of its image on the retina

(Fogassi et al., 1992; Galletti et al., 1993). If similar

principles could be applied to the transformation of

vestibular input from a head to a body-centred reference

frame through regulatory neck afferents (Manzoni, 2005),

we could expect that neck rotation modulates response

gain at earlier stages and induces shifts in preferred

directions at last stages of the processing. Thus, if the

recorded reticular units project to cerebellar structures, a

‘‘gain field’’ modulation of their vestibular responses by

neck rotation can be expected. On the other hand,

although many of the recorded neurons were located

within the cerebellar projecting paramedian reticular

nucleus (Newman and Ginsberg, 1992; see Ito, 1984 for

ref.), a similar number of units were clearly located

outside this area. Moreover the paramedian reticular

nucleus of rat shows a high concentration of

reticulospinal neurons (Reed et al., 2008). Thus, it is

unlikely that the gain field behaviour documented in the

presented study merely represents a precerebellar stage

of processing.

An alternative hypothesis may be that the framework of

a serial processing of information in neurons projecting to

the cerebellum, cerebellar cortex, RF and spinal cord is

not sufficient for an appropriate tuning of the VS reflex.

Other upstream structures may be involved in the

transformation, and the cerebellum could be responsible

for their correct integration by generating appropriate

predictive signals related to the final motor response

(Ebner and Palasar, 2008). Thus, ‘‘gain fields’’ in

reticulospinal neurons, following a spinal integration step,

could give rise to a directional tuning at motoneuronal

levels. Upstream structures possibly involved in the

neural transformation could be the motor layers of

the superior colliculus, which are connected to the

reticulospinal neurons (Isa and Sasaki, 2002; Perkins et

al., 2009), and receive information from the cerebellum

(May et al., 1990). This modifies the discharge and

responsiveness of about half of the collicular neurons

(Niemi-Junkola and Westby, 2000), which are known to

integrate not only visual and acoustic (Meredith et al.,

1992; Stein, 1998), but also neck (Abrahams and Rose,

1975; Nagy and Corneil, 2010) and labyrinthine

information (Maeda et al., 1979; Bacskai et al., 2002).

Moreover, collicular activation during eye movements is

tuned by the relative trunk-to-head position (Nagy and

Corneil, 2010). Further investigation is requested in order

to verify whether collicular neurons show labyrinthine

responses tuned by neck rotation and whether reticular

neurons with ‘‘gain field’’ type responses are actually

projecting to the spinal cord or to the cerebellum.

Other possible targets for reticular units

Finally we wish to underline that medullar reticular cells are

the primary source of projections to vestibular efferent

A C E

B D F

Fig. 7. Changes in the gain and direction of the responses to wobble stimuli induced by trunk rotation in reticular units with different gain behaviour.

(A) Gain changes. The response gain to CW (filled symbols) and CCW rotation (open symbols) has been plotted as a function of the corresponding

trunk displacement for 8 reticular units with positive gain trend. Data are expressed in % with respect to the average of all the gain values obtained at

the different amplitudes of body rotation analysed. In (A)–(D) dotted lines represent the regression equation for all the plotted points (see text for

further details). (B) Directional changes have been shown for the same units illustrated in (A) by using identical symbols. Values of 0� in ordinate

corresponds to the response direction evaluated in the control position. (C) Gain changes. The response gain to CW rotation has been plotted as a

function of the corresponding trunk rotation for 3 reticular units with negative gain trend. Data are expressed in percentage (%) of the average of all

the gain values obtained at the different amplitudes of body rotation analysed. (D) Directional changes of the same units illustrated in (C) have been

plotted by using the same symbols. Values of 0� in ordinate correspond to the response direction evaluated in the control position. (E) Gain changes.

The response gain to CW (filled symbols) and CCW rotation (open symbols) has been plotted as a function of the corresponding trunk displacement

for 5 ‘‘tuned’’ reticular units. Data are expressed in % of the average of all the gain values obtained at different amplitudes of body rotation analysed.

(F) Directional changes have been plotted for the same units illustrated in (E) by using the same symbols and lines. Zero values on abscissa

represent the response direction obtained in the control position (with the exception of the unit indicated with the open circles that was not

responsive in this position). In (A)–(F) positive and negative values on abscissa refer to ipsilateral and contralateral trunk rotations, respectively.

60 M. Barresi et al. / Neuroscience 224 (2012) 48–62

neurons (Metts et al., 2006). Theoretically the units we

recorded from may project to the efferent system, thus

modifying the vestibular afferent discharge by a feed-

back vestibular signal modulated by a tonic neck input.

This loop does exist (Plotnik et al., 2002), but its effects

on the afferent activity show up when the velocity of

rotation exceeds 80�/s, a value that is much higher with

respect to peak velocities in pitch and roll utilized in the

present experiments.

CONCLUSIONS

Our findings indicate that RF units take part in the process

that transforms the reference frame for vestibular signals

from head to body centred frames. However, the pattern

of modulation of their vestibular response by neck

rotation was different from that observed both in the

cerebellum and in the spinal motoneurons, thus further

integration may occur at premotoneuronal level.

As a consequence, the role exerted by the cerebellum

in the integration of vestibular and neck has to be

re-considered.

Acknowledgements—The present investigation was supported

by grants of the Italian Space Agency (ASI, DCMC project and

Grant I/R/335/02) and by the University of Pisa. We thank P.

Orsini and G. Montanari for their valuable technical assistance

and G. Bertolini for animal care. The help of Dr. E. Santarcangelo

in editing and revising the manuscript is gratefully acknowledged.

REFERENCES

Abrahams VC, Rose PK (1975) Projections of extraocular, neck

muscle, and retinal afferents to superior colliculus in the cat: their

M. Barresi et al. / Neuroscience 224 (2012) 48–62 61

connections to cells of origin of tectospinal tract. J Neurophysiol

38:10–18.

Angelaki DE (1991) Dynamic polarization vector of spatially tuned

neurons. IEEE Trans Biomed Eng 38:1053–1060.

Angelaki DE (1992) Two-dimensional coding of linear acceleration

and the angular velocity sensitivity of the otolith system. Biol

Cybern 67:511–521.

Bacskai T, Szekely G, Matesz C (2002) Ascending and descending

projections of the lateral vestibular nucleus in the rat. Acta Biol

Hung 53:7–21.

Barresi M, Grasso C, Li Volsi G, Manzoni D (2010) Cerebellar control

of input–output coupling within vestibulospinal reflexes: role of the

lateral vestibular nucleus and the medullary reticular formation.

Front Neurosci Conference Abstract: The cerebellum: from

neurons to higher control and cognition 2010. http://dx.doi.org/

10.3389/conf.fnins.83.00013.

Bolton PS, Goto T, Schor RH, Wilson VJ, Yamagata Y, Yates BJ

(1992) Response of pontomedullary reticulospinal neurons to

vestibular stimuli in vertical planes. Role in vertical vestibulospinal

reflexes of decerebrate cat. J Neurophysiol 67:639–647.

Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD, Marsden

CD (1993) Postural electromyographic responses in the arm and

leg following galvanic vestibular stimulation in man. Exp Brain Res

94:143–151.

Brotchie PR, Andersen RA, Lawrence HS, Goodman S (1995) Head

position signals used by parietal neurons to encode locations of

visual stimuli. Nature 375:232–235.

Bush GA, Perachio AA, Angelaki DE (1993) Encoding of head

acceleration in vestibular neurons: I. Spatiotemporal response

properties to linear acceleration. J Neurophysiol 69:2039–2055.

Cenciarini M, Peterka R (2006) Stimulus-dependent changes in

vestibular contribution to human postural control. J Neurophysiol

95:2733–2750.

Coulter JD, Mergner T, Pompeiano O (1976) Effect of static tilt on

cervical spinoreticular tract neurons. J Neurophysiol 39:

45–62.

Denoth F, Margherini PC, Pompeiano O, Stanojevic M (1979)

Responses of Purkinje cells of the cerebellar vermis to neck

and macular vestibular inputs. Pflugers Arch 381:87–98.

Duhamel JR, Bremmer F, Behamed S, Graf W (1997) Spatial

invariance of visual receptive fields in parietal cortex neurons.

Nature 389:845–848.

Ebner TJ, Palasar S (2008) Cerebellum predicts the future motor

state. Cerebellum 7:583–588.

Ezure K, Wilson VJ (1983) Dynamics of the neck-to-forelimb reflexes

in the decerebrate cat. J Neurophysiol 50:688–695.

Fitzpatrick R, Burke D, Gandevia SC (1994) Task-dependent reflex

responses and movement illusions evoked by galvanic vestibular

stimulation in standing humans. J Physiol 478:363–372.

Fogassi L, Gallese V, di Pellegrino G, Fadiga L, Gentilucci M, Luppino

G, Matelli M, Pedotti A, Rizzolatti G (1992) Space coding by

premotor cortex. Exp Brain Res 89:686–690.

Galletti C, Battaglia PP, Fattori P (1993) Parietal neurons encoding

spatial locations in craniotopic coordinates. Exp Brain Res

96:221–229.

Grant G, Oscarsson O, Rosen I (1966) Functional organization of the

spino-reticulo-cerebellar path with identification of its spinal

components. Exp Brain Res 1:306–319.

Igarashi M, Watanabe T, Maxian PM (1970) Dynamic equilibrium in

squirrel monkeys after unilateral and bilateral labyrinthectomy.

Acta Otolaryngol 69:247–253.

Isa T, Sasaki S (2002) Brainstem control of head movements during

orienting: organization of the premotor circuits. Prog Neurobiol

66:205–241.

Ito M (1984) The cerebellum and neural control. New York: Raven

Press.

Kammermeier S, Klein J, Buttner U (2009) Vestibular–neck

interaction in cerebellar patients. Ann N Y Acad Sci

1164:394–399.

Kasper J, Schor RH, Wilson VJ (1988) Response of vestibular

neurons to head rotations in vertical planes: II. Response to the

neck stimulation and vestibular–neck interaction. J Neurophysiol

60:1765–1778.

Kleine JF, Guan Y, Kipiani E, Glonti L, Hoshi M, Buttner U (2004)

Trunk position influences vestibular responses of fastigial

nucleus neurons in the alert monkey. J Neurophysiol 91:

2090–2100.

Lackner JR, Dizio P, Jeka J, Horak F, Krebs D, Rabin E (1999)

Precision contact of the fingertip reduces postural sway of

individual with bilateral vestibular loss. Exp Brain Res

126:459–466.

Ladpli R, Brodal A (1968) Experimental studies of commissural and

reticular formation projections from the vestibular nuclei in the cat.

Brain Res 8:65–96.

Lindsay KW, Roberts TD, Rosemberg JR (1976) Asymmetric tonic

labyrinth reflexes and their interaction with neck reflexes in

decerebrate cat. J Physiol 261:583–601.

Lund S, Broberg C (1983) Effects of different head positions on

postural sway induced by a reproducible vestibular error signal.

Acta Physiol Scand 117:307–309.

Maeda M, Shibazaki T, Yoshida K (1979) Labyrinthine and visual

inputs to the superior colliculus neurons. Prog Brain Res

50:735–743.

Magnus R (1928) Korpestellung. Berlin: Springer.

Manzoni D (2005) The cerebellum may implement the appropriate

coupling of sensory inputs and motor responses: evidence from

vestibular physiology. Cerebellum 4:178–188.

Manzoni D, Pompeiano O, Srivastava UC, Stampacchia G (1983)

Responses of forelimb extensors to sinusoidal stimulation of

macular labyrinth and neck receptors. Arch Ital Biol 121:

205–214.

Manzoni D, Pompeiano O, Andre P (1998) Neck influences on the

spatial properties of vestibulospinal reflex in decerebrate cats:

role of the cerebellar anterior vermis. J Vest Res 8:283–297.

Manzoni D, Pompeiano O, Bruschini L, Andre P (1999) Neck input

modifies the reference frame for coding labyrinthine signals in the

cerebellar vermis: a cellular analysis. Neuroscience 3:

1095–1107.

May PJ, Hartwich-Young R, Nelson J, Sparks DL, Porter JD (1990)

Cerebellotectal pathways in the macaque: implications for

collicular generation of saccades. Neuroscience 36:305–324.

Meredith MA, Wallace MT, Stein BE (1992) Visual, auditory and

somatosensory convergence in output neurons of the cat superior

colliculus: multisensory properties of the tecto-reticular–spinal

projection. Exp Brain Res 88:181–186.

Mergner T, Huber W, Becker W (1997) Vestibular–neck interaction

and transformation of sensory coordinates. J Vest Res

7:347–367.

Mergner T, Siebold C, Schweigart G, Becker W (1991) Human

perception of horizontal trunk and head rotation in space during

vestibular and neck stimulation. Exp Brain Res 85:389–404.

Nagy B, Corneil BD (2010) Representation of horizontal head-on-

body position in the primate superior colliculus. J Neurophysiol

103:858–874.

Metts BA, Kaufman GD, Perachio AA (2006) Polysynaptic inputs to

vestibular efferent neurons as revealed by viral transneuronal

tracing. Exp Brain Res 172:261–274.

Newman DB, Ginsberg CY (1992) Brainstem reticular nuclei that

project to the cerebellum in rats: a retrograde tracer study. Brain

Behav Evol 39:194.

Niemi-Junkola UJ, Westby GW (2000) Cerebellar Output exerts

spatially organized influence on neural responses in rat superior

colliculus. Neuroscience 97:565–573.

Perkins E, Warren S, May PJ (2009) The mesencephalic reticular

formation as a conduit for primate collicular gaze control: tectal

inputs to neurons targeting the spinal cord and the medulla. Anat

Rec (Hoboken) 292:1162–1181.

Peterson BW, Abzug C (1975) Properties of projections from

vestibular nuclei to medial reticular formation in the cat. J

Neurophysiol 38:1421–1435.

Petrovicky P (1976) Distribution and organization of spino-reticular

afferents in the brain stem of rat. J Hirnforsch 17:127–135.

62 M. Barresi et al. / Neuroscience 224 (2012) 48–62

Plotnik M, Marlinski V, Goldberg JM (2002) Reflection of efferent

activity in rotational responses chinchilla vestibular afferents. J

Neurophysiol 88:1234–1244.

Pompeiano O (1967) Functional organization of the cerebellum

protections to the spinal cord. Prog Brain Res 25:282–321.

Pompeiano O, Andre P, Manzoni D (1997) Spatiotemporal response

properties of cerebellar Purkinje cells to animal displacements: a

population analysis. Neuroscience 81:609–626.

Reed WR, Shum-Siu A, Magnuson DS (2008) Reticulospinal

pathways in the ventrolateral funiculus with terminations in the

cervical and lumbar enlargements of the adult rat spinal cord.

Neuroscience 24:505–517.

Roberts TDM (1978) Neurophysiology of postural mechanisms. 2nd

ed. London: Butterworths.

Roll R, Velay LJ, Roll JP (1991) Eye and neck proprioceptive

messages contribute to the spatial coding of retinal input in

visually oriented activities. Exp Brain Res 85:423–431.

Salinas E, Abbott LF (1995) Transfer of coded information from

sensory to motor networks. J Neurosci 15:6461–6474.

Shaikh AG, Meng H, Angelaki DE (2004) Multiple reference frames

for motion in the primate cerebellum. J Neurosci 24:4491–4497.

Schor RH, Angelaki DE (1992) The algebra of neuronal response

vectors. Ann N Y Acad Sci 656:190–204.

Schor RH, Miller AD, Tomko DL (1984) Responses to head tilt in cat

central vestibular neurons: I. Direction of maximum sensitivity. J

Neurophysiol 51:136–146.

Stein BE (1998) Neural mechanisms for synthesizing sensory

information and producing adaptive behaviours. Exp Brain Res

123:124–135.

Uchino Y, Sato H, Kushiro K, Zakir MM, Isu N (2000) Canal and

otolith inputs to single vestibular neurons in cats. Arch Ital Biol

138:3–13.

Von Holst E, Mittelstaedt H (1950) Das Reafferenzprinzip

(Wechselwirkung zwischen zentralnervensystem und

peripherie). Naturwissenschaften 37:464–476.

Welgampola MS, Colebatch JC (2001) Vestibulospinal reflexes:

quantitative effects on sensory feedback and postural task. Exp

Brain Res 139:345–353.

Wilson VJ, Schor RH, Suzuki I, Park BR (1986) Spatial organization

of neck and vestibular reflexes acting on the forelimbs of the

decerebrate cat. J Neurophysiol 55:514–526.

(Accepted 7 August 2012)(Available online 14 August 2012)