ORIGINAL PAPER
R. S. Sadka Æ Z. Wollberg
Response properties of auditory activated cells in the occipital cortex
of the blind mole rat: an electrophysiological study
Received: 6 August 2003 / Revised: 19 January 2004 / Accepted: 29 January 2004 / Published online: 17 March 2004

Springer-Verlag 2004
Abstract Previous studies have demonstrated that de-
spite its blindness, the subterranean blind mole rat
(Spalax ehrenbergi) possesses a noticeable lateral
geniculate nucleus and a typical cyto-architectural
occipital cortex that are reciprocally connected. These
two areas, as revealed by the metabolic tracer 2-de-
oxyglucose, are activated by auditory stimuli. Using
single unit recordings, we show that about 57% of 325
cells located within the occipital cortex of anesthetized
mole rats responded to at least one of the following
auditory stimuli — white noise, pure tones, clicks, and
amplitude modulated tones — with the latter two being
the most effective. About 85% of cells driven by either
contralateral or ipsilateral stimulation also responded
to binaural stimulation; about 13% responded only to
binaural stimulation; and 2% were driven exclusively
by contralateral stimulation. Comparing responsiveness
and response strength to these three modes of stimu-
lation revealed a contralateral predominance. Mean
latency (±SD) of ipsilateral and contralateral re-
sponses were 48.5±32.6 ms and 33.5±9.4 ms, respec-
tively. Characteristic frequencies could be divided into
two distinct subgroups ranging between 80 and 125 Hz
and between 2,500 and 4,400 Hz, corresponding to the
most intensive spectral components of the vibratory
intraspecific communication signals and airborne
vocalizations.
Keywords Auditory responses Æ Mole rat, Spalax
ehrenbergi Æ Occipital cortex
Abbreviations BMF: best modulation frequency Æ CF:
characteristic frequency Æ 2-DG: 2-deoxyglucose Æ
dLGN: dorsal lateral geniculate nucleus Æ IC: inferior
colliculus Æ LGN: lateral geniculate nucleus Æ OC:
occipital cortex Æ MTF: modulation transfer
function Æ SAM: sinusoidally amplitude modulation Æ
SC: superior colliculus
Introduction
Recent years have revealed a growing interest in cross-
modal neuroplasticity — the recruitment of brain areas
normally used for the processing of one sensory
modality, for the processing of another sensory modal-
ity. Compensatory neuronal reorganization and the
diversion of one sensory modality to targets of another
sensory modality have been demonstrated in various
animal models deprived of a particular sensory input
either congenitally or experimentally (e.g., Asanuma
et al. 1988; Asanuma and Stanfield 1990; Schlaggar and
O’Leary 1991; Bronchti et al. 1992; Toldi et al. 1994a,
1994b, 1996; Rauschecker and Korte 1993; Rauschecker
and Kniepert 1994;Yaka et al. 1999, 2000; Izraeli et al.
2002).
Cross-modal reorganization in animal models has
been induced, in most cases, by experimental proce-
dures. For example, following removal of the lateral
geniculate nucleus (LGN) and the superior colliculus
(SC) in newborn Syrian hamsters and ferrets and the
creation of an alternative target for the visual input by
partial deafferentation of the somatosensory or audi-
tory thalamic nuclei, visual fibers that had lost their
original terminal space formed new connections in the
‘abandoned’ thalamic nuclei. It has been shown that
neural responses evoked by visual stimuli in the
somatosensory or auditory cortex of such neonatally
operated hamsters and ferrets were similar to those of
visual cortex cells in sighted hamsters and ferrets and
that the surgically induced pathway could mediate vi-
sual pattern discrimination (e.g., Sur et al. 1988, 1990;
Frost 1990; Pallas et al. 1990; Roe et al. 1990, 1992;
Pallas and Sur 1994; Frost et al. 2000; von Melchner
et al. 2000). The activation of primary visual areas
R. S. Sadka Æ Z. Wollberg (&)
Department of Zoology, George S. Wise Faculty of Life Sciences,
Tel Aviv University, Tel Aviv, 69978, Israel
E-mail: zviw@tauex.tau.ac.il
Fax: +972-3-6407271
J Comp Physiol A (2004) 190: 403–413
DOI 10.1007/s00359-004-0506-7

by other sensory modalities found in blind animal
models is consistent with findings in early blind
human subjects, as revealed electrophysiologically and by
non-invasive brain imaging methods (e.g., Veraart et al.
1990; Uhl et al. 1991; Alho et al. 1993; Kujala et al. 1995,
2000; Sadato et al. 1996, 1998; De Volder et al. 1997;
Roder et al. 2000).
In our current animal model, the subterranean
blind mole rat (Spalax ehrenbergi, also known as
Nannospalax ehrenbergi), we take advantage of blind-
ness that is natural and congenital. Possession of a
dorsal lateral geniculate nucleus (dLGN) and occipital
cortex (OC) (Bronchti et al. 1989, 2002) render this
rodent a good model system for studying cross-modal
compensation. Indeed, earlier functional mapping
using 2-deoxyglucose (2-DG) has revealed that in the
blind mole rat these visual areas can be activated by
auditory stimuli (Bronchti et al. 1989; Heil et al. 1991;
Bronchti et al. 2002). The major source of this audi-
tory input has been shown to be the inferior colliculus
(IC), which in addition to all typical auditory targets,
also projects to the dLGN (Doron and Wollberg
1994). Preliminary electrophysiological experiments
also disclosed single unit responses to various auditory
stimuli in the striate cortex (Heil et al. 1991). The
present study describes and characterizes single unit
responses to auditory stimuli in the blind mole rat’s
OC, expanding our previous functional mapping and
projection tracing studies that indicated a takeover by
auditory input of the OC in this naturally blind
rodent.
Materials and methods
Animal preparation and surgical procedures
A total of 24 mole rats (S. ehrenbergi) of both sexes were captured
in the northern part of the Negev desert in Israel. In the laboratory,
animals were housed individually in separate cages under constant
temperature (25±2C) and light/dark (14/10 h) illumination sche-
dule, with food (vegetables and laboratory rat chow) provided ad
libitum.
All surgical procedures and electrophysiological recordings
were conducted under deep anesthesia. Because metabolic, respi-
ratory and heart rates of the blind mole rat are normally very low
(Arieli et al. 1977; Storier et al. 1981), attaining prolonged and
stable anesthesia was problematic. Moreover, inter-individual
differences in sensitivity to the anesthetic drugs were remarkable.
We achieved a moderately stable and non-fatal anesthesia in most
cases by an initial i.m. administration of a mixture of ketamin
hydrochloride (27 mg kg

1
body mass) and xylazine hydrochlo-
ride 2% (0.3 ml kg

1
body mass). Recordings lasted several
hours. Anesthesia was maintained throughout the experiment by
subcutaneous administration of supplementary doses of the mix-
ture as required by the different individuals (half of the initial
dose, every 30–60 min). Anesthetized animals were restrained in a
specially designed stereotaxic apparatus that left the ears unob-
structed. The posterior skull was exposed by removing the over-
lying muscle, and a 5 mm diameter circular craniotomy was made
unilaterally over the OC leaving the dura intact. To prevent
desiccation of the exposed brain it was covered with paraffin oil.
The animal was then placed in a suspended double-walled, sound-
attenuation chamber (IAC 1203A) lined with sound-damping
foam to reduce echo. All the electrophysiological recordings were
conducted within this chamber.
Stimulation and recording procedures
Auditory stimuli consisted of 0.2 ms clicks (presented individually
or in trains of four successive 0.2-ms clicks), broadband white noise
(0.02–20 kHz) produced by a noise generator (Bru¨el and Kjaer type
1405; Naerum, Denmark) and pure tones generated by a voltage-
controlled oscillator (Wavetek type 136, San Diego, USA ). The
latter were presented in linear ascending order in two separate
series: 16 consecutive steps ranging from 100 to 1,300 Hz; and 64
consecutive steps ranging from 1,030 to 13,900 Hz. Noise and tones
were shaped into 200-ms bursts with 20 ms rise and fall time, using
a custom-made shaping unit. To simulate the pulsed periodic
vibratory signals of the blind mole rat (Rado et al. 1987) we also
used sinusoidally amplitude-modulated (SAM) tones shaped by a
two-quadrant multiplier (responding to the positive voltages and
chopping the negative voltages of the symmetrical sinusoidal
modulating signal). The carrier frequency was 100 Hz, corre-
sponding to the dominant frequency component in the natural
seismic vibrations of the mole rat. Modulation frequencies varied
between 2 and 64 Hz. Stimuli strength ranged between 80 and
90 dB SPL (unless indicated differently). Amplitudes were con-
trolled by means of a custom-made power amplifier and a manually
operated attenuator (HP350D Hewlett-Packard).
Stimuli were delivered at a repetition rate of 0.5 Hz through
pre-calibrated earphones (Azden, New York, USA) coated and
isolated from the animal’s ears by vibration absorbing material.
Signals were presented binaurally or monaurally, both contralat-
eral and ipsilateral to the recording side. Sound pressure levels (in
dB re. 20 lPa) were monitored by a calibrated condenser micro-
phone (Bru¨el & Kjaer 4134) and a sound-level meter (Bru¨el & Kjaer
2209) serially connected to a 1/3-octave filter set (Bru¨el & Kjaer
1616). Frequency response of the entire sound-delivery system was
fairly flat (±4 dB) throughout the frequency range of 70 Hz to
20 kHz.
Single unit activity was extracellularly recorded using glass-
coated platinum-iridium microelectrodes that were advanced
through the dura by a calibrated and remotely controlled stepping
motor (resolution: 4.5 lm/step; Sodeco-Saia, model AMA92,
Switzerland). Of 55 penetrations, 51 were perpendicular to the
brain surface. Most of the cells included in this study were de-
tected through their spontaneous activity. However, we also
occasionally used auditory ‘search’ stimuli. Neuronal activity was
a.c. amplified, filtered (Digitimer Neurolog System; Digitimer
Research Instrumentation, England), monitored for shape and
size and discriminated from background activity by a window
discriminator (WPI type 121; World Precision Instruments, USA).
Discriminated spikes were digitized and displayed on-line as dot
rasters. Field-evoked potentials used to delineate the occipital
responsive area and to distinguish it from the auditory cortex
were picked up by the same microelectrodes, a.c. amplified, fil-
tered (bandwidth 0.05–5 kHz), averaged (RC Electronics, Ore.,
USA) and displayed on-line. Both cellular and field-evoked
potentials were stored for off-line analyses on a personal com-
puter. During the experiment, recording sites were stereotactically
referred to the mid-line and the anterior border of the straight
part of the transverse sinus.
At the end of the electrophysiological experiments, animals
were anesthetized with fluothane and perfused transcardially with
0.9% saline solution followed by 10% neutral buffered formalin.
The brains were then gently removed and deep frozen (
40C).
Assignment of electrode penetration tracks and recording sites was
based on the calibrated microelectrode driver and coronal Nissl-
stained frozen sections. The midline and the splenium of the corpus
callosum served as internal references for orientation along the
mediolateral and anteroposterior axes, respectively. In order to
compare our results with earlier work on the auditory and
somatosensory systems of this species, we adopted the schematic
brain views and coordinate system of Necker et al. (1992) and
404