J Neurophysiol 112: 328 –339, 2014.
First published April 30, 2014; doi:10.1152/jn.00016.2014.
Sensory processing and corollary discharge effects in posterior caudal lobe
Purkinje cells in a weakly electric mormyrid fish
Karina Alviña and Nathaniel B. Sawtell
Department of Neuroscience, Columbia University, New York, New York
Submitted 6 January 2014; accepted in final form 25 April 2014
cerebellum; corollary discharge; synaptic plasticity; electric fish
THE BRAINS OF MOST VERTEBRATES contain both a cerebellum and
sensory processing structures similar to the cerebellum in terms
of their evolution, development, patterns of gene expression,
and circuitry (Bell 2002; Bell et al. 2008). Studies of cerebellum-like structures involved in processing electrosensory information in fish suggest functional similarities as well. These
structures, which include the dorsal octavolateral nucleus of
elasmobranch fish and the electrosensory lobe (ELL) of weakly
electric mormyrid and gymnotid fish, appear to act as adaptive
sensory processors in which diverse sensory and motor signals
conveyed by a granule cell-parallel fiber system cancel out
predictable patterns of sensory input via mechanisms of associative synaptic plasticity (Bell 2001; Bodznick et al. 1999; Bol
et al. 2011). This mode of operation resembles Marr-Albus and
adaptive filter models of the cerebellum (Albus 1971; Dean et
al. 2010; Fujita 1982; Marr 1969). Moreover, studies of the
mormyrid ELL have provided a detailed mechanistic account
of how motor corollary discharge signals related to the fish’s
electric organ discharge (EOD) are used to predict and cancel
the electrosensory consequences of the EOD (Bell 1981; Bell
et al. 1997b; Kennedy et al. 2014; Roberts and Bell 2000).
Intriguingly, this role for ELL circuitry closely resembles
Address for reprint requests and other correspondence: N. B. Sawtell,
Columbia Univ., Dept. of Neuroscience, HHSC RM 501, 701 W 168th St.,
New York, NY 10032 (e-mail: [email protected]).
328
predictive or forward model functions hypothesized for the
mammalian cerebellum (Ebner and Pasalar 2008; Izawa et al.
2012; Miall et al. 1993; Wolpert et al. 1998). However, in
addition to these similarities, there are also important differences between cerebellum-like structures and the cerebellum,
most notably the presence of climbing fiber (CF) input from the
inferior olive exclusively to the latter. Given that many aspects
of mammalian cerebellar function remain unresolved, examining the similarities and differences between the cerebellum and
cerebellum-like structures may provide a useful perspective on
cerebellar function in general.
The nervous system of mormyrid fish offers a remarkable
opportunity for such a comparison. Anatomical studies have
shown that the posterior caudal lobe (LCp) of the mormyrid
cerebellum receives the same sources of mossy fiber input and
projects to the same target structures as ELL (Bell et al. 1981;
Campbell et al. 2007), i.e., there is a region of the mormyrid
cerebellum proper that is closely associated and arranged in
parallel with ELL. Granule cells that send parallel fibers to the
molecular layer of ELL are located in the same external
granule cell mass, known as the eminentia granularis posterior
(EGp), as the granule cells that send parallel fibers to LCp
(Fig. 1A). Given that mossy fibers terminate widely within
EGp, it is likely that ELL and LCp receive similar patterns of
mossy fiber input (Bell et al. 1981; Campbell et al. 2007).
These include corollary discharge input associated with the
EOD motor command, electrosensory information, and proprioceptive signals (Bell et al. 1992; Sawtell 2010). LCp and ELL
also have similar projection patterns. Glutamatergic efferent
cells of ELL project to higher stages of electrosensory processing in the midbrain—the preeminential nucleus and the lateral
toral nucleus. Glutamatergic efferent cells of LCp, equivalent
to the deep cerebellar nuclear cells of higher vertebrates,
project to these same structures (Bell et al. 1981; Campbell et
al. 2007). LCp, but not ELL, also projects to two small nuclei,
the rostral and caudal paraventricular nuclei, about which little
is known. Additional similarities between LCp and ELL include several classes of interneurons found in both structures.
These are the Golgi cells and unipolar brush cells located in
EGp and the inhibitory molecular layer interneurons located in
both LCp and ELL molecular layers (Campbell et al. 2007;
Kennedy et al. 2014).
It has been suggested that the presence of a distinct CF input
from the inferior olive is the key feature that distinguishes the
cerebellum proper from cerebellum-like structures (Bell 2002;
Bell et al. 2008). This is the case for LCp versus ELL. While
LCp Purkinje cells integrate parallel fiber input from EGp with
CF input from the inferior olive, ELL principal cells integrate
parallel fiber input from EGp with somatotopically organized
0022-3077/14 Copyright © 2014 the American Physiological Society
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
Alviña K, Sawtell NB. Sensory processing and corollary discharge
effects in posterior caudal lobe Purkinje cells in a weakly electric
mormyrid fish. J Neurophysiol 112: 328 –339, 2014. First published
April 30, 2014; doi:10.1152/jn.00016.2014.—Although it has been
suggested that the cerebellum functions to predict the sensory consequences of motor commands, how such predictions are implemented
in cerebellar circuitry remains largely unknown. A detailed and
relatively complete account of predictive mechanisms has emerged
from studies of cerebellum-like sensory structures in fish, suggesting
that comparisons of the cerebellum and cerebellum-like structures
may be useful. Here we characterize electrophysiological response
properties of Purkinje cells in a region of the cerebellum proper of
weakly electric mormyrid fish, the posterior caudal lobe (LCp), which
receives the same mossy fiber inputs and projects to the same target
structures as the electrosensory lobe (ELL), a well-studied cerebellum-like structure. We describe patterns of simple spike and climbing
fiber activation in LCp Purkinje cells in response to motor corollary
discharge, electrosensory, and proprioceptive inputs and provide evidence for two functionally distinct Purkinje cell subtypes within LCp.
Protocols that induce rapid associative plasticity in ELL fail to induce
plasticity in LCp, suggesting differences in the adaptive functions of
the two structures. Similarities and differences between LCp and ELL
are discussed in light of these results.
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
tipolar and exhibit more irregular dendritic branching patterns
(Fig. 1C) (Campbell et al. 2007; Zhang and Han 2007).
In contrast to the great deal of information available about in
vivo response properties and plasticity in ELL and EGp,
knowledge of LCp is based on a single previous in vivo study
that reported briefly on electrosensory and corollary discharge
responses in Purkinje cells (Bell et al. 1992). The present study
provides the first detailed description of in vivo response
properties of LCp Purkinje cells with attention to both simple
spike and CF responses. Our findings include prominent and
temporally diverse corollary discharge responses in Purkinje
cells, evidence for two functionally distinct subregions within
LCp, and apparent differences in the conditions required to
induce associative plasticity in LCp versus ELL.
A
B
329
C
Fig. 1. Inputs, outputs, and circuitry of posterior caudal lobe (LCp) vs.
electrosensory lobe (ELL). A: a Nissl-stained transverse section showing LCp
and ELL. Mossy fibers convey a variety of sensory and motor signals to
eminentia granularis posterior (EGp). EGp granule cells (GCs, filled circles) in
turn send parallel fibers to LCp and ELL molecular layers. Dorsal (dPC) and
multipolar (mPC) Purkinje cells (gray) also receive a climbing fiber input from
the inferior olive, while Purkinje-like medium ganglion (MG) cells of ELL
(gray) receive input from peripheral electroreceptors. Glutamatergic efferent
neurons of ELL (orange) and LCp (cyan) project in parallel to higher stages of
electrosensory processing. Whether or not dorsal and multipolar Purkinje cells
converge onto the same efferent cells, as depicted, is not known. rpv and cpv,
rostral and caudal paraventricular nuclei. B: sagittal section showing a fanshaped Purkinje cell located in the dorsal region of LCp filled with biocytin
during intracellular recording. C: sagittal section showing a multipolar Purkinje cell located in the medial region of LCp filled with biocytin during
intracellular recording.
input from peripheral electroreceptors. Given the distinctive
properties of CFs and the critical role they play in most
theoretical accounts of cerebellar function, this difference may
be extremely important. In vitro studies have revealed several
additional differences between Purkinje cells in LCp and Purkinje-like medium ganglion (MG) cells in ELL. These include
a prominent NMDA receptor-mediated component of parallel
fiber synaptic transmission in MG cells but not Purkinje cells
and dendritic spikes that appear to be primarily sodium mediated in MG cells and calcium mediated in Purkinje cells (Grant
et al. 1998; Zhang et al. 2010; Zhang and Han 2007).
To date, knowledge of LCp comes mainly from in vitro
electrophysiological and anatomical studies. In vitro studies
have shown that basic intrinsic and synaptic properties of LCp
Purkinje cells are similar in many respects to those of mammalian Purkinje cells (Zhang and Han 2007). Anatomical
studies have provided evidence for distinct Purkinje cell subtypes within LCp (Campbell et al. 2007; Zhang and Han 2007).
Unlike most cerebella, LCp lacks a Purkinje cell layer; rather,
Purkinje cell bodies are scattered throughout the molecular
region. Purkinje cells located in the most dorsal part of the
molecular layer have fan-shaped dendrites that extend toward
the dorsal surface of LCp (Fig. 1B), while those located in the
medial and ventral portions of the molecular region are mul-
Experimental preparation. All experiments performed in this study
were approved by the Columbia University Institutional Animal Care
and Use Committee and adhere to the American Physiological Society’s “Guiding Principles in the Care and Use of Vertebrate Animals
in Research and Training.” Mormyrid fish (7–12 cm in length) of the
species Gnathonemus petersii were used in these experiments. Surgical procedures to expose EGp for recording were identical to those
described previously (Sawtell 2010). Briefly, fish were anesthetized
(MS-222, 1:25,000) and held against a foam pad. Skin on the dorsal
surface of the head was removed, and a long-lasting local anesthetic
(0.75% bupivacaine) was applied to the wound margins. A plastic rod
was cemented to the anterior portion of the skull to hold the head
rigid. The posterior portion of the skull was removed, and the
underlying valvula cerebelli was reflected laterally to expose EGp and
the molecular layer of LCp. At the end of the surgery, a paralytic,
gallamine triethiodide (Flaxedil), was given (⬃20 ␮g/cm of body
length), the anesthetic was removed, and aerated tank water was
passed over the fish’s gills for respiration. Paralysis blocks the effect
of electromotoneurons on the electric organ, preventing the EOD, but
the motor command signal that would normally elicit an EOD continues to be emitted by the electromotoneurons at a variable rate of
2–5 Hz. The timing of the EOD motor command can be measured
precisely (see below), and the central effects of electric organ corollary discharge (EOCD) inputs can be observed in isolation from the
electrosensory input that would normally result from the EOD. Methods for electrosensory stimulation and for generating controlled
movements of the tail were the same as those described previously
(Bell 1982; Bell and Grant 1992; Sawtell 2010).
Electrophysiology. The EOD motor command signal was recorded
with an electrode placed over the electric organ in the tail. The
command signal is the synchronized volley of electromotoneurons
that would normally elicit an EOD in the absence of neuromuscular
blockade. The command signal lasts ⬃3 ms and consists of a small
negative wave followed by three larger biphasic waves. The latencies
of central corollary discharge or command-evoked responses were
measured with respect to the negative peak of the first large biphasic
wave in the command signal. EGp and the LCp molecular layer can
be directly visualized after reflecting the overlying cerebellar valvula.
Extracellular recordings from LCp Purkinje cells were made with
glass microelectrodes filled with 2 M NaCl. Methods for in vivo
whole cell current-clamp recordings were the same as those described
previously (Sawtell 2010). Briefly, electrodes (9 –15 M⍀) were filled
with an internal solution containing (in mM) 122 K-gluconate, 7 KCl,
10 HEPES, 0.4 Na2GTP, 4 MgATP, and 0.5 EGTA, with 0.5%
biocytin (pH 7.2, 280 –290 mosM). No correction was made for liquid
junction potentials. Only cells with stable membrane potentials more
hyperpolarized than ⫺45 mV and access resistance ⬍100 M⍀ were
analyzed. Membrane potentials were filtered at 3–10 kHz and digi-
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
METHODS
330
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
RESULTS
Basic electrophysiological properties of LCp Purkinje cells.
LCp Purkinje cell recordings were characterized by the presence of two distinct all-or-none events that differed both in
their waveforms (Fig. 2, A and C) and frequency of occurrence
[simple spikes: 25.9 ⫾ 12.8 Hz (n ⫽ 100 extracellular record-
A
Fig. 2. Electrophysiological properties of LCp
Purkinje cells. A: extracellular trace from a
Purkinje cell illustrating simple spike (SS) and
climbing fiber (CF, arrow) response. Histogram
shows a pause in simple spike firing following a
CF response. B: overlaid intracellular traces
showing CF responses evoked by microstimulation (left) and those occurring spontaneously
(right) in the same Purkinje cell. C: intracellular
traces from 3 different Purkinje cells showing
the varied appearance of simple and CF responses. Top: a putative somatic recording.
Middle and bottom: putative dendritic recordings. Scatterplot in inset shows amplitudes of
simple and CF responses for a population of
intracellularly recorded Purkinje cells (n ⫽ 61).
D: in addition to spontaneous simple and CF
responses (filled arrow), some Purkinje cells
exhibited a third broad all-or-none event (open
arrow), presumably a dendritic spike, that could
be evoked by strong depolarizing current injection. E: overlaid intracellular traces and smoothed
simple spike firing rates (25-ms Gaussian kernel;
dotted lines are SE) triggered on the occurrence of
a putative dendritic spike (top, open arrow) or a
CF response (bottom, filled arrow) in the same
Purkinje cell. Putative dendritic spikes were invariably preceded by a gradual depolarization and
acceleration of simple spike firing and followed by
a pause in firing (top, upper trace). This distinctive
pattern was not associated with CF responses
(bottom, upper trace).
ings) and 18.9 ⫾ 13.2 (n ⫽ 55 whole cell recordings); CF
responses: 0.47 ⫾ 0.34 (n ⫽ 55 whole cell recordings)]. The
occurrence of a CF response was typically accompanied by a
pause in simple spike firing (Fig. 2, A and E, bottom). The
appearance of CF responses in our in vivo recordings is
somewhat different from those observed in mammalian Purkinje cells but similar to those reported previously for
mormyrid Purkinje cells in vitro (de Ruiter et al. 2006; Han and
Bell 2003; Zhang and Han 2007). These studies have shown
that CF activation evokes all-or-none excitatory postsynaptic
potentials (EPSPs) in mormyrid Purkinje cells without obvious
spikelike components (de Ruiter et al. 2006). All-or-none
responses similar or identical to spontaneous CF responses
could be evoked by microstimulation near the brain surface just
anterior to LCp where the tract in which the axons of inferior
olive neurons travel (Campbell et al. 2007), but not of surrounding regions (Fig. 2B). This observation is consistent with
CF responses indeed being due to CF activation. Simple spike,
but not CF response, amplitude varied widely in our whole cell
recordings, without any obvious relation to recording quality
(Fig. 2C). Most likely, this variation in simple spike amplitude
is due to differences in recording location (somatic vs. dendritic) and to the failure of simple spikes to propagate into LCp
Purkinje cell dendrites, as shown previously for Purkinje cells
in other species (e.g., Llinas and Sugimori 1980a).
In addition to simple spikes and CF responses, we also
observed a third distinct type of spike that could be evoked by
strong membrane potential depolarization. These spikes were
much broader than simple spikes, and unlike CF responses they
B
D
C
E
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
tized at 20 kHz (CED power1401 hardware and Spike2 software;
Cambridge Electronics Design, Cambridge, UK).
Histology. After recording, fish were deeply anesthetized with a
concentrated solution of MS-222 (1:10,000) and transcardially perfused with a teleost Ringer solution followed by a fixative consisting
of 2% paraformaldehyde and 2% glutaraldehyde or 4% paraformaldehyde in 0.1 M phosphate buffer. The brains were postfixed, cryoprotected with 20% sucrose, and sectioned at 50 ␮m on a cryostat.
Sections were reacted with avidin-biotin complex and diaminobenzidine or a streptavidin-conjugated fluorescent dye to reveal the biocytin-filled cells.
Data analysis and statistics. Data were analyzed off-line with
Spike2 and MATLAB (MathWorks, Natick, MA). Data are expressed
as means ⫾ SD, unless otherwise noted. Paired and unpaired Student’s t-tests were used to test for statistical significance, as noted.
Differences were judged to be significant at P ⬍ 0.05. Only recordings from Purkinje cells, as judged by the presence of two distinct
spike waveforms one much more frequent than the other, were
included in the analysis. Unless stated otherwise, analysis of EOCD
responses used only data from EOD commands separated by 200 ms
or greater. Simple spike modulations in response to the EOD command were calculated as the difference between maximum evoked
firing rate and the minimum evoked firing rate divided by the mean
rate.
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
were always preceded by a strong membrane potential depolarization and high rates of simple spike firing (Fig. 2, D and
E). Presumably these are dendritic spikes, as described in
mormyrid Purkinje cells in vitro (Han and Bell 2003; Zhang
and Han 2007). Consistent with a possible dendritic origin, we
A
C
D
E
F
G
observed that these spikes were most easily evoked via current
injections in putative dendritic recordings in which simple
spikes were small (Fig. 2D). Finally, we occasionally observed
distinctive firing patterns that resembled the trimodal pattern
that has been described in Purkinje cells in other systems
(Llinas and Sugimori 1980b; Womack and Khodakhah 2004;
Zhang and Han 2007). These consisted of a gradual acceleration of simple spike firing terminated by a dendritic spike and
a subsequent pause in simple spike firing (Fig. 2E, top).
Electric organ corollary discharge responses in LCp Purkinje cells. EOCD responses have been well-studied in ELL,
where they have been shown to play a variety of functions,
including the generation of negative images of self-generated
electrosensory input (Bell 2001). Although EOCD responses in
LCp Purkinje cells were described briefly in a previous study
(Bell et al. 1992), they have not been thoroughly characterized.
As in previous studies, we take advantage of an awake preparation in which fish continue to emit the motor command to
discharge their electric organ but the EOD itself is blocked by
neuromuscular paralysis, allowing EOCD responses, i.e., neural activity in sensory areas that is time-locked to the EOD
motor command, to be studied in isolation from electrosensory
effects. Patterns of simple spike and CF responses to the EOD
motor command were analyzed for 55 Purkinje cells recorded
intracellularly. Nearly all cells exhibited some EOCD response, though the strength of such responses varied widely
across cells. Figure 3A shows eight Purkinje cells selected for
their prominent simple spike modulations. Two features of
these responses are notable. First, a wide variety of relationships were observed between temporal patterns of simple spike
firing and CF responses. In some cases simple spikes and CF
responses were roughly antiphasic (Fig. 3A, top left), in others
they were roughly in phase (Fig. 3A, top right), while some
cells exhibited strong simple spike modulations without any
modulation of CF responses (Fig. 3A, bottom). Second,
whereas the EOD motor command itself is brief and highly
stereotyped, effects of the command on Purkinje cells were
substantially more delayed and temporally diverse. EOCD
responses in LCp Purkinje cells are also considerably more
delayed and diverse than EOCD responses observed previously
in most granule cells in EGp (Fig. 3B) (Kennedy et al. 2014)
and in MG cells in mormyromast zones of ELL (Fig. 3C) (Bell
Fig. 3. Electric organ corollary discharge (EOCD) responses in LCp Purkinje
cells. A: electric organ discharge (EOD) command-evoked simple spike and
CF response rates smoothed with a 25-ms Gaussian filter for 8 different
Purkinje cells recorded intracellularly. Dotted lines are SE. Rasters correspond
to the uppermost example in each row. Small black circles are simple spikes,
and large gray circles are CFs. B, top: smoothed firing rate (25-ms Gaussian
filter) of a typical EGp granule cell in response to the EOD motor command.
Bottom: overlaid membrane potential (Vm) traces showing command-evoked
depolarization and spiking in the same example granule cell. By comparison,
responses of Purkinje cells are far more delayed and temporally diverse. C:
EOD command-evoked narrow spike rates (25-ms Gaussian filter) overlaid for
10 different MG cells recorded in the mormyromast zones of ELL. Cells were
selected at random from a large number of MG cell recordings from previous
studies. Despite receiving similar parallel fiber input as LCp Purkinje cells,
EOCD responses are less diverse and delayed in MG cells. D: individual trace
from an intracellularly recorded Purkinje cell, illustrating subthreshold membrane potential modulation in relation to the EOD command. Filled arrow
indicates CF response. Bottom trace is the electromotoneuron volley (EMN)
measured near the tail. E–G: average command-evoked subthreshold responses
for 3 Purkinje cells; E shows same cell as in D. Gray outlines are SE.
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
B
331
332
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
A
B
C
D
et al. 1997a; Mohr et al. 2003a; Sawtell et al. 2007). Diverse,
delayed, and long-lasting subthreshold responses to the EOD
command were also observed in intracellular recordings from
Purkinje cells (Fig. 3, D–G). In some Purkinje cells subthreshold responses resembled a continuous oscillation locked to the
EOD command (Fig. 3, D and E), while in others commandevoked EPSPs and inhibitory postsynaptic potentials (IPSPs)
with sharp onsets were evident (Fig. 3, F and G).
Variation in the strength of EOCD responses was strongly
associated with recording depth. Purkinje cells encountered
near the surface of LCp, which is directly visible after reflection of the overlying valvula cerebelli, tended to have weaker
EOCD responses than those encountered deeper in LCp (Fig. 4,
A and B). Figure 4C shows the strength of simple spike
modulation for 100 extracellularly recorded Purkinje cells in
three fish plotted versus recording depth. Given their superficial locations, weakly responding cells probably correspond to
the dorsal Purkinje cells described in previous anatomical and
in vitro studies of LCp (Campbell et al. 2007; Zhang and Han
2007). Consistent with this, morphologically identified dorsal
Purkinje cells had smaller simple spike modulations in response to EOCD inputs than morphologically identified multipolar Purkinje cells [0.96 ⫾ 0.36 (n ⫽ 11) for dorsal cells vs.
2.58 ⫾ 0.86 (n ⫽ 9) for multipolar cells; P ⬍ 0.0001, Student’s
t-test]. These results suggest that dorsal and multipolar Purkinje cells are functionally distinct.
Finally, we noted that in some Purkinje cells accelerations in
the rate of EOD motor commands emitted by the fish were
associated with strong modulations of simple spike firing. Such
accelerations can be evoked by sensory stimuli, the so-called
novelty response (Post and von der Emde 1999), but also
occurred spontaneously (Fig. 5A, top). These modulations did
A
D
B
C
Fig. 5. Effects of EOD command rate on LCp Purkinje cell responses.
A: spontaneous EOD command rate accelerations. Top: times of occurrence of
individual EOD commands. Bottom: instantaneous command rate. B: EOD
command-evoked simple spike firing rates (25-ms Gaussian filter) for 2
different Purkinje cells (black vs. gray trace) recorded in the same fish as the
command intervals shown in A. Averages were constructed with commands
separated by ⬎150 ms. C: opposite effects of command rate increases on
simple spike firing for the same 2 Purkinje cells. D, top: intracellular traces
from a different Purkinje cell illustrating that depolarization and increases in
simple spike firing often precede increases in EOD command rate (arrows).
Middle: times of occurrence of individual EOD commands. Bottom: instantaneous command rate.
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
Fig. 4. Strength of LCp Purkinje cell EOCD responses varies with recording
depth. A: command-evoked simple spike firing rate (25-ms Gaussian filter;
dotted lines are SE) from a Purkinje cell recorded extracellularly near the
surface of LCp (⬃80 ␮m deep); “mod” indicates the modulation of simple
spike firing by the EOD motor command (see METHODS). B: command-evoked
simple spike firing rate from a Purkinje cell recorded extracellularly deeper
within LCp (⬃380 ␮m deep). C: command-evoked simple spike modulations
plotted vs. recording depth for n ⫽ 100 extracellularly recorded Purkinje cells
in 3 fish. D: average simple spike firing rate plotted vs. recording depth for the
cells shown in C.
not appear to be the result of simple temporal summation of
EOCD responses at high command rates, as cells with similar
EOCD responses measured for normal EOD command intervals (⬎150 ms) (Fig. 5B) could have opposite responses to
accelerations (Fig. 5C). Also, it was evident from intracellular
recordings that in some cells depolarization consistently preceded the accelerations in EOD command rate (Fig. 5D) and
hence could not result from simple temporal summation. Such
responses were observed in roughly half of the Purkinje cells
we recorded.
Electrosensory responses in LCp Purkinje cells. Two main
classes of electroreceptors project to ELL: ampullary receptors
subserving passive electrolocation and mormyromast receptors
subserving active electrolocation (Bell 1986b). While ampullary receptors are sensitive to low-frequency external fields,
e.g., those emitted by other animals, and project to the ventrolateral zone of ELL, mormyromast receptors are sensitive to
small changes in amplitude and waveform of the fish’s own
EOD and project to the medial and dorsolateral zones of ELL.
All three zones of ELL project to the midbrain preeminential
nucleus, which, in turn, provides mossy fiber input to EGp
(Bell and Szabo 1986).
We first tested responses of Purkinje cells to low-frequency
electrosensory stimuli designed to selectively engage ampul-
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
A
C
cells and to those recorded in superficially located Purkinje
cells (Fig. 4). Finally, all five Purkinje cells with responses to
low-frequency electrosensory stimuli that were morphologically identified were dorsal Purkinje cells.
We next tested responses of Purkinje cells to stimuli designed to engage mormyromast receptors involved in active
electrolocation. These consisted of brief pulses delivered
shortly after the EOD command, mimicking the fish’s own
EOD. EOD mimics could be delivered either locally to restricted regions of the body surface to map receptive fields or
globally to activate the entire receptor surface. In contrast to
ELL (Bell and Grant 1992; Metzen et al. 2008; Sawtell and
Williams 2008) or preeminential nucleus neurons (Sawtell et
al. 2005; von der Emde and Bell 1996), LCp Purkinje cells did
not respond strongly to local stimuli and appeared to lack
well-delineated electrosensory receptive fields. As shown in
Fig. 7, we did observe changes in simple spike firing and CF
response patterns when we turned on a global EOD mimic
locked to the EOD motor command. Although the EOD mimic
activates both mormyromast and ampullary receptors (Bell and
Russell 1978), we think it is likely that at least some of the
responses we observed are related to activation of mormyromast receptors. If responses to EOD mimics were due exclusively to ampullary receptor activation, they should be strongest in superficial, putative dorsal Purkinje cells with weak
EOCD responses. This did not appear to be the case. As can be
seen in Fig. 7 and Fig. 9A, some Purkinje cells with responses
to EOD mimics also exhibited strong EOCD responses.
D
Fig. 6. Effects of low-frequency electrosensory
stimuli (ES) on LCp Purkinje cell responses. A,
top: intracellular trace showing subthreshold, simple spike, and CF response modulations evoked by
a global sinusoidal ES (middle). Bottom: times of
occurrence of individual EOD commands. B: simple spike and CF response rates (25-ms Gaussian
filter) in relation to the phase of a 0.5-Hz global
sinusoidal ES presented at 3 different stimulus
intensities to the same cell. The phase of the ES is
shown at bottom of B–D. Note that CF responses
are modulated at lower stimulus intensities than
simple spikes. C: simple spike and CF response
rates (25-ms Gaussian filter) in relation to the
phase of a sinusoidal ES presented at 5 different
frequencies to the same cell (different from that
shown in B). D: simple spike and CF response
rates (25-ms Gaussian filter) in relation to the
phase of a sinusoidal ES delivered via a local
dipole positioned at 4 different locations on the
body indicated by numbers in the schematic at top
(different cell from those shown in B and C).
Different locations of the local ES resulted in
different phase relationships to the stimulus,
though in each case simple spikes and CF responses were antiphasic.
B
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
lary receptors. Such stimuli clearly modulated simple spikes
and CF responses in putative dorsal Purkinje cells (Fig. 6A).
Such stimuli modulated simple spikes and CF responses at
intensities far below those required to activate mormyromast
receptors and within the range shown to be effective for
activating ampullary receptors (Bell 1982) (Fig. 6B). Several
features of these responses were notable. Simple spike modulations and CF responses were strongest at low frequencies,
typically peaking at ⬃1 Hz and declining sharply at higher
frequencies (Fig. 6C). With local stimulation it was often
possible to evoke different responses at different skin positions, with some positions yielding no response (Fig. 6D).
Finally, modulations of simple spikes and CF responses were
typically antiphasic (Fig. 6).
Several observations suggest that responses to low-frequency electrosensory stimuli are specific to dorsal Purkinje
cells. First, such responses were found almost exclusively at
superficial depths. This observation was confirmed in a subset
of experiments in which recording depth and responsiveness to
low-frequency electrosensory stimulation were tracked for all
recorded Purkinje cells. Fifteen of thirty-four extracellularly
recorded Purkinje cells exhibited simple and/or CF response
modulations in response to low-frequency electrosensory stimuli. Responsive cells were recorded at more superficial locations than nonresponsive cells (253 ⫾ 121.2 ␮m vs. 450.7 ⫾
133.3 ␮m; P ⬍ 0.0001, Student’s t-test). Low-frequency responsive cells also exhibited weak EOCD responses (modulation index: 0.66 ⫾ 0.32, n ⫽ 13 whole cell recordings), similar
to those observed in morphologically identified dorsal Purkinje
333
334
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
A
B
Proprioceptive responses in LCp Purkinje cells. Mossy
fibers originating from the spinal cord and brain stem convey
proprioceptive information about the posture and movements
of the fish to EGp (Bell et al. 1981, 1992). Previous studies
have suggested that such information is used by ELL neurons
to cancel out self-generated changes in electrosensory input
due to movements of the electric organ, located in the tail,
relative to electroreceptors on the body surface (Bell et al.
1981; Sawtell 2010; Sawtell and Williams 2008). Purkinje
cells commonly exhibited membrane potential, simple spike,
and CF response modulations in response to passive sinusoidal
displacements of the tail (0.1–1 Hz) (Fig. 8). A variety of
response patterns were observed (Fig. 8, A and B), consistent
with the variety of responses to tail position, velocity, and
speed observed previously in mossy fibers (Sawtell 2010).
Modulations of simple spike firing and CF responses were
often, though not exclusively, antiphasic (Fig. 8).
Modulations of simple spike and CF responses by tail
movements were sometimes observed in the same Purkinje
cells that exhibited simple spike and CF response modulations
in response to low-frequency electrosensory stimuli (Fig. 8D;
n ⫽ 8), suggestive of multimodal integration at the level of
both Purkinje cells and the inferior olive. We also observed
that Purkinje cells exhibiting clear modulation of simple spike
and/or CF responses to tail displacements exhibited weaker
EOCD responses than those that did not respond to tail movements [modulation index: 0.99 ⫾ 0.57 vs. 2.89 ⫾ 0.91 (n ⫽ 30
and 12 whole cell recordings, respectively); P ⬍ 0.00001,
Student’s t-test]. Of 30 cells that responded to tail movements,
9 were morphologically identified as dorsal and 3 as multipolar
Purkinje cells. Of 12 Purkinje cells that showed no response to
tail movements, 6 were morphologically identified as multipolar Purkinje cells.
LCp Purkinje cells do not exhibit rapid associative plasticity
in vivo. Motivated by similarities between negative images
described previously in ELL and mammalian cerebellar plasticity and learning (see DISCUSSION), we attempted to induce
associative plasticity of EOCD and proprioceptive inputs in
LCp Purkinje cells by pairing the EOD motor command or a
proprioceptive stimulus with activation of CFs. First, we compared EOCD responses before, during, and after pairing with a
A
B
C
D
Fig. 8. Effects of passive tail movements on LCp Purkinje cell responses. A and
B: 2 example Purkinje cells showing simple spike and CF response rates
(25-ms Gaussian filter) in relation to the phase of a sinusoidal tail displacement
(⫾20°) presented at 3 different frequencies to each cell. Note that simple spike
and CF responses are roughly antiphasic. C: intracellular recording showing
the average subthreshold response (Vm, middle) and CF responses (raster, top)
in response to a sinusoidal tail displacement (bottom). D: example cell with
simple spike and CF responses to both sinusoidal tail displacement (⫾20°)
(top) and a low-frequency ES (bottom).
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
Fig. 7. Effects of an EOD mimic on LCp Purkinje cell responses. A, left: EOD
command-evoked simple spike and CF response rates (25-ms Gaussian filter;
dotted lines are SE) for an example Purkinje cell recorded intracellularly.
Right: effects of a global EOD mimic delivered 4.5 ms after the command
(arrow). B: same display for a second example Purkinje cell. Rasters correspond to the cell shown in A.
global EOD mimic locked to the EOD motor command (Fig. 9A).
For the example Purkinje cell shown in Fig. 9A, the global
EOD mimic caused a clear shift in the latency of commandevoked CF responses during pairing (Fig. 9A, bottom). In
contrast to results obtained previously in ELL neurons (Bell
1981; Bell et al. 1997a; Bell and Grant 1992), no temporally
specific change was apparent after 45 min of pairing (Fig. 9A,
top). Similar experiments were performed for 13 cells, with
pairing durations ranging from 5 to 45 min. Although we
sometimes observed overall shifts in membrane potential, no
temporally specific changes in EOD command-evoked subthreshold or simple spike responses were observed after pairing
in these experiments. A second series of experiments were
performed in which a single dendritic spike evoked by intracellular current injection was paired at a fixed delay after the
EOD command for between 2 and 10 min (Fig. 9B, bottom).
These pairings also failed to reveal temporally specific changes
in EOCD responses (Fig. 9B, top; n ⫽ 11 cells). In a third
series of experiments we compared subthreshold and simple
spike responses to the EOD command before, during, and after
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
A
335
Purkinje cells, independent of changes in synaptic transmission
(Cerminara and Rawson 2004; Colin et al. 1980; Demer et al.
1985; Mathews et al. 2012; McKay et al. 2007; Montarolo et
al. 1982; Savio and Tempia 1985).
Finally, two additional sets of pairing experiments were
conducted to test the possibility that proprioceptive inputs to
LCp Purkinje cells exhibit plasticity. Simple spike responses to
sinusoidal tail displacements were compared before, during,
and after 5–20 min of pairing with either an electrosensory
stimulus (200-ms square pulse; Fig. 10A, n ⫽ 9 cells) or
electrical microstimulation evoking CF responses (Fig. 10B,
n ⫽ 3 cells) at a fixed phase of the tail displacement. Electrosensory stimuli effectively drove CF responses in dorsal Pur-
B
B
C
Fig. 9. LCp Purkinje cells do not exhibit rapid associative plasticity of EOCD
responses. A, top: average subthreshold EOCD responses before (black) and
after (gray) pairing the EOD command with a global EOD mimic (bottom) for
45 min. Although this cell exhibited an apparent increase in CF response rate
after pairing, such changes were not consistently observed. B: same display as
in A, but pairing was conducted by injecting current to evoke a dendritic spike
at a fixed delay after the command (bottom). Pairing duration for the cell
shown was 7 min. C: EOD command-evoked simple spike and CF response
rates (25-ms Gaussian filter) for an extracellularly recorded Purkinje cell
before (pre), during (pairing w CF stim), and at 2 times after (post) pairing the
EOD command with a CF response evoked by microstimulation for 45 min.
Inset: the evoked CF waveform. Although the overall rate of simple spike
discharge was reduced, there was little change in the temporal pattern of simple
spike firing.
pairing with microstimulation-evoked CF responses (Fig. 9C,
inset). Compared with pairings using the global EOD mimic
described above, microstimulation provided a much greater
degree of control over the rate and timing of CF responses
relative to the command. Pairings lasted between 3 and 40 min
and were conducted in 12 cells. Despite the stronger engagement of CF responses in these experiments, temporally specific
changes in command-evoked subthreshold or simple spike
responses were not consistently observed after pairing (Fig.
9C). We did, however, observe overall reductions in simple
spike firing after pairing in some experiments (Fig. 9C, compare pre vs. post). Such reductions could also be induced
simply by increasing the rate of CF stimulation independent of
the EOD command (data not shown) and hence are unlikely to
reflect associative plasticity. These observations are consistent
with results in mammalian Purkinje cells showing that changes
in the rate of CF activation alter intrinsic excitability of
Fig. 10. LCp Purkinje cells do not exhibit rapid associative plasticity of
proprioceptive responses. A: simple spike and CF response rates (25-ms
Gaussian filter) in relation to the phase of a sinusoidal tail displacement (⫾20°)
before (top), during (middle), and after (bottom) pairing with an ES (200-ms
square pulse, inset) delivered at a fixed phase of the displacement for 10 min.
The ES reliably evoked CF responses during pairing but did not cause a change
in simple spike firing after pairing. B: simple spike and CF response rates
(25-ms Gaussian filter) in relation to the phase of a sinusoidal tail displacement
before (top), during (middle), and after (bottom) pairing with CF responses
evoked by microstimulation (arrows, inset) at a fixed phase of the displacement
for 10 min. Scale bar for inset, 20 ms.
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
A
336
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
kinje cells, presumably via activation on low-frequency electroreceptors. Despite large changes in the pattern of CF responses during pairing, patterns of simple spike firing in
relation to tail displacements were largely unchanged after
pairing.
DISCUSSION
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
This study provides an initial characterization of the response properties of Purkinje cells in a region of the mormyrid
cerebellum that is closely associated with the electrosensory
system and has afferent and efferent connections similar to an
extremely well-studied cerebellum-like structure—the ELL.
One firm conclusion of our study is that functional differences
exist between previously described dorsal and multipolar Purkinje cells located in different subregions of LCp (Campbell et
al. 2007; Zhang and Han 2007). Strong responses to lowfrequency electrosensory and proprioceptive stimuli were common in putative dorsal Purkinje cells (located superficially in
LCp) but rare or absent in putative multipolar Purkinje cells
(located deeper in LCp). EOCD responses, on the other hand,
were stronger in putative multipolar Purkinje cells. In some
cases, putative multipolar Purkinje cells with prominent EOCD
responses could be shown to respond to a global EOD mimic
with changes in both simple spike firing and CF responses (Fig.
7). A similar pattern is observed in regions of ELL involved in
passive versus active electrosensory processing, i.e., EOCD
responses are much stronger in the latter (Bell 1982; Bell et al.
1992; Bell and Grant 1992). Hence, as in ELL, distinct LCp
subregions may be dedicated to passive versus active electrosensory processing.
Differences in EOCD responses we observed between dorsal
and multipolar Purkinje cells are also consistent with a previous report that EOCD field potentials recorded in medial
portions of EGp are much smaller than those recorded more
laterally in EGp (Bell et al. 1992). EGp granule cells located
medially are expected to send parallel fibers to the dorsal
region of LCp, while granule cells located laterally are expected to project to the deeper portion of LCp as well as to ELL
(Fig. 1A). Hence a possible explanation for the functional
differences we observed is that information from ampullary
and mormyromast regions of ELL remains segregated at the
level of the preeminential nucleus and is returned to different
subregions of EGp. This hypothesis could be tested with tracer
injections into different zones of ELL and LCp. Responses of
putative dorsal Purkinje cells to low-frequency electrosensory
stimuli were more prominent than those of putative multipolar
Purkinje cells to EOD mimics. This is surprising given that
most of ELL and most of the preeminential nucleus are
dedicated to active electrolocation. More complex or naturalistic electrosensory stimuli may be required to strongly engage
multipolar Purkinje cells. Consistent with this possibility, a
study in weakly electric gymnotid fish found that moving
electrosensory stimuli were more effective than stationary ones
in modulating activity in cerebellar neurons (Bastian 1975).
The specific role(s) of LCp in electrosensory processing is
unknown. One possibility is that LCp plays a relatively direct
role in some form of electrosensory-guided behavior. One such
behavior is the electromotor response, in which fish use electrosensory information to stabilize their position relative to a
moving refuge (Cowan and Fortune 2007; Rose and Canfield
1993). This behavior is similar to well-studied optomotor
responses. Adaptive modification of the gain of the optomotor
response in larval zebrafish appears to be cerebellum dependent (Ahrens et al. 2012). Alternatively, LCp may play a more
general role in electrosensory processing. The fact that LCp
projects (in parallel with ELL) mainly to electrosensory processing regions rather than motor centers seems to suggest the
latter. Behavioral studies coupled with lesions of LCp could
begin to address this issue.
Although the histological structure and connections of various regions of the cerebellum and the in vitro electrophysiological properties of various cerebellar cell types have been
well-studied in mormyrid fish (Campbell et al. 2007; Han et al.
2007; Han and Bell 2003; Meek et al. 2008; Nieuwenhuys and
Nicholson 1969; Zhang and Han 2007), very little is known
about signals conveyed by CFs (Bell et al. 1992; Russell and
Bell 1978). Our study provides some initial insights into this
important issue. CF responses in LCp Purkinje cells were
observed at specific delays relative to the EOD command and
brief EOD mimic pulses and at specific phases in relation to
low-frequency electrosensory stimuli and passive tail movements. CF responses in some Purkinje cells were modulated in
relation to both low-frequency electrosensory and proprioceptive signals or to both EOCD and electrosensory signals,
suggesting multimodal convergence of sensory and motor
signals onto individual inferior olive neurons. Our results
clearly indicate that signals conveyed by CFs to LCp are more
diverse and complex than those conveyed by electroreceptors
to ELL. Studies aimed at determining how CFs are engaged by
more natural patterns of electrosensory input or in the context
of electrosensory-guided behavior are needed in order to generate more specific hypotheses regarding LCp function.
Comparisons with ELL. Intracellular and extracellular recordings from LCp Purkinje cells presented here, together with
past in vitro studies of mormyrid Purkinje cells (de Ruiter et al.
2006; Han and Bell 2003; Zhang et al. 2012; Zhang and Han
2007), suggest that the basic electrophysiological properties of
LCp Purkinje cells are quite similar to those described for
Purkinje cells in other species. Similar to LCp Purkinje cells,
the Purkinje-like MG cells of ELL also exhibit multiple types
of all-or-none events. Narrow spikes are frequent and of axonal
origin and convey the main output of MG cells, while broad
spikes are infrequent and of dendritic origin and are key
triggers for associative plasticity at parallel fiber synapses (Bell
et al. 1997b; Engelmann et al. 2008; Grant et al. 1998).
However, unlike CF responses in LCp Purkinje cells, MG cell
broad spikes are not related to a CF input and can be readily
evoked by somatic current injections at thresholds only modestly greater than those evoking axonal spikes. MG cell broad
spikes also appear to differ from dendritic spikes in LCp
Purkinje cells in that they are not associated with a trimodal
firing pattern and are sodium spikes rather than calcium spikes
(Grant et al. 1998; Zhang et al. 2010; Zhang and Han 2007).
LCp Purkinje cells exhibited subthreshold and simple spike
responses to the EOD motor command, electrosensory stimuli,
and proprioceptive stimuli. Presumably such responses are
mediated by mossy fiber-granule cell inputs. EOCD and proprioceptive responses have also been well-documented for
ELL neurons, highlighting the fact that the same mossy fiber
inputs are processed in parallel by a cerebellum-like circuit and
a region of the cerebellum proper. EOCD responses provide an
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
proaches that have revealed plastic changes in the responses of
mammalian Purkinje cells in vivo (e.g., Jirenhed et al. 2007;
Jirenhed and Hesslow 2011). These experiments took advantage of the fact that the fish’s own EOD command provides
strong and natural activation of EOCD mossy fibers and evokes
strong depolarizations in many LCp Purkinje cells, presumably
mediated, at least in part, by parallel fiber synapses (Fig. 3,
B–E). Similarly, passive tail displacements were used to provide strong and reasonably naturalistic engagement of proprioceptive mossy fibers. Activation of CFs at particular delays
after the EOD command or particular phases of tail displacements was achieved via electrical microstimulation or with
electrosensory stimuli. Such pairings failed to induce consistent changes in membrane potential and/or simple spike firing
that were specific to the timing or phase of the CF responses
during pairing. Although we sometimes observed overall
changes in membrane potential and simple spike firing rates
after pairing, such changes were nonspecific and were most
prominent in experiments in which electrical microstimulation
was used to evoke CF responses at rates substantially higher
than those normally observed. These effects are likely due to
changes in excitability rather than associative synaptic plasticity. Such excitability changes are a known consequence of
changing the rate of CF input (Cerminara and Rawson 2004;
Colin et al. 1980; Demer et al. 1985; Mathews et al. 2012;
McKay et al. 2007; Montarolo et al. 1982; Savio and Tempia
1985). Although our failure to induce plasticity is not surprising on its own, it does contrast with ELL, where rapid associative plasticity is readily induced in vivo in a preparation
identical to that used in the present study.
Several explanations can be offered for why we failed to
induce associative plasticity of EOCD or proprioceptive inputs.
First, such plasticity may not exist in LCp. We believe this is
unlikely in light of previous in vitro studies of Purkinje cells in
the central lobes of the mormyrid cerebellum demonstrating
plasticity of parallel fiber synapses induced by pairing with CF
stimulation (Han et al. 2007). In addition, preliminary observations suggest that similar plasticity can be induced in vitro in
LCp Purkinje cells (V. Han, unpublished observations). A
second possibility is that plasticity driven by coactivation of
parallel fiber and climbing input exists in LCp but our experimental protocols were simply insufficient to induce such
plasticity. Negative images in ELL neurons can be induced
very rapidly, over a timescale of just a few minutes of pairing
(Bell 1982; Bell et al. 1997a; Bell and Grant 1992; Sawtell and
Williams 2008). Although a variety of different protocols have
been used to induce plasticity in Purkinje cell in vivo, some
studies, including eyelid conditioning paradigms discussed
above (Jirenhed et al. 2007; Jirenhed and Hesslow 2011),
suggest that the induction of in vivo plasticity in the cerebellum
may require longer-duration pairing periods. Different requirements for plasticity induction in LCp versus ELL may be
related to previously described differences in parallel fiber
synaptic transmission, i.e., the presence of a prominent NMDA
receptor-mediated component of parallel fiber-evoked responses in MG but not LCp Purkinje cells (Grant et al. 1998;
Zhang and Han 2007).
The possibility of different timescales of plasticity in LCp
versus ELL suggests an interpretation of the observation that
EOCD responses in LCp Purkinje cells are more delayed and
diverse than in MG cells, despite the fact that both cell types
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
interesting point of comparison because they have been wellstudied both in ELL and in EGp. The EOD motor command,
like the EOD itself, is brief and extremely stereotyped. In
contrast, EOCD responses in putative multipolar Purkinje cells
were often delayed and long-lasting and showed considerable
diversity in their temporal patterns. Recent recordings and
modeling of EOCD responses in EGp granule cells suggest that
Purkinje cell responses we observed would not be expected
based on simple summation of granule cell EOCD input. The
large majority of granule cells are active at short delays after
the EOD command (Fig. 3B), with only a small fraction of cells
active at longer delays (Kennedy et al. 2014).
Experimental and theoretical studies of negative image formation and sensory cancellation in ELL have shown that
anti-Hebbian spike timing-dependent plasticity at parallel fiber-MG cell synapses, together with an array of granule cells
active at a range of different delays after the EOD motor
command, provides a mechanism for generating diverse and
delayed EOCD responses (Bell 1981; Bell et al. 1997b; Kennedy et al. 2014; Roberts and Bell 2000). The possibility that
similar mechanisms are involved in generating diverse and
delayed EOCD responses observed in multipolar Purkinje cells
is discussed below. Other mechanisms, besides plasticity,
could also be responsible for shaping EOCD responses in LCp.
LCp Purkinje cells receive inhibitory inputs from molecular
layer interneurons and also inhibit each other, as shown by in
vitro paired recordings (Zhang et al. 2012).
To the extent to which they are understood, adaptive processes in the cerebellum appear similar in many respects to
negative images and sensory cancellation described in ELL
(Bell et al. 2008). In ELL, the pairing of parallel fiber signals,
e.g., EOCD or proprioceptive input, with peripheral sensory
input results in such signals eliciting a predictive reduction in
principal cell activity, as described above for EOCD signals. In
the cerebellum, pairing of parallel fiber signals with CF input
leads to such signals eliciting a reduction in the firing of
Purkinje cells. If the CFs convey some type of sensory signal,
gated through the inferior olive, then the parallel fiber signals
that are paired with the CFs and that predict their occurrence
will reduce Purkinje cell activity.
For example, adaptively timed pauses in Purkinje cell firing
observed in the context of eyelid conditioning are similar in
numerous respects to temporally specific negative images in
ELL neurons (Jirenhed et al. 2007; Jirenhed and Hesslow
2011). A leading hypothesis regarding the mechanisms underlying such changes in Purkinje cell responses is similar to that
described above for temporally specific negative images in
ELL. Temporal patterns of simple spike firing in Purkinje cells
are sculpted via CF-induced plasticity at parallel fiber-Purkinje
cell synapses acting on an array of granule cell inputs with a
variety of temporal activity patterns in relation to the conditioned stimulus (Medina et al. 2000; Medina and Mauk 2000).
Extending these ideas to LCp, it could be hypothesized that
temporal patterns of EOCD response in Purkinje cells are
sculpted via CF-induced plasticity acting on an array of differently timed EOCD input conveyed by parallel fibers.
We addressed this hypothesis by experimentally manipulating the relationship between mossy fiber and CF inputs. Our
approach was similar to that used previously to reveal negative
images in ELL neurons (Bell 1981; Bell et al. 1993; Sawtell
2010; Sawtell and Williams 2008) and similar also to ap-
337
338
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
ACKNOWLEDGMENTS
Present address of K. Alviña: Albert Einstein College of Medicine, Dept. of
Neuroscience, 1410 Pelham Pkwy S., Rm. 703, Bronx, NY 10461.
GRANTS
This work was supported by grants from the National Science Foundation
(1025849), the National Institute of Neurological Disorders and Stroke (NS075023), the Alfred P. Sloan Foundation, and the McKnight Endowment Fund
for Neuroscience to N. B. Sawtell.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
AUTHOR CONTRIBUTIONS
Author contributions: K.A. and N.B.S. performed experiments; K.A. and
N.B.S. analyzed data; K.A. and N.B.S. edited and revised manuscript; K.A.
and N.B.S. approved final version of manuscript; N.B.S. conception and design
of research; N.B.S. interpreted results of experiments; N.B.S. prepared figures;
N.B.S. drafted manuscript.
REFERENCES
Ahrens MB, Li JM, Orger MB, Robson DN, Schier AF, Engert F,
Portugues R. Brain-wide neuronal dynamics during motor adaptation in
zebrafish. Nature 485: 471– 477, 2012.
Albus JS. A theory of cerebellar function. Math Biosci 10: 25– 61, 1971.
Bastian J. Receptive fields of cerebellar cells receiving exteroceptive input in
a gymnotid fish. J Neurophysiol 38: 285–300, 1975.
Bell CC. An efference copy modified by reafferent input. Science 214:
450 – 453, 1981.
Bell CC. Properties of a modifiable efference copy in electric fish. J Neurophysiol 47: 1043–1056, 1982.
Bell CC. Duration of plastic change in a modifiable efference copy. Brain Res
369: 29 –36, 1986a.
Bell CC. Electroreception in mormyrid fish: central physiology. In: Electroreception, edited by Bullock TH, Heiligenberg W. New YorkWiley, 1986b,
p. 423– 452.
Bell CC. Memory-based expectations in electrosensory systems. Curr Opin
Neurobiol 11: 481– 487, 2001.
Bell CC. Evolution of cerebellum-like structures. Brain Behav Evol 59:
312–326, 2002.
Bell CC, Caputi A, Grant K. Physiology and plasticity of morphologically
identified cells in the mormyrid electrosensory lobe. J Neurosci 17: 6409 –
6422, 1997a.
Bell CC, Caputi A, Grant K, Serrier J. Storage of a sensory pattern by
anti-Hebbian synaptic plasticity in an electric fish. Proc Natl Acad Sci USA
90: 4650 – 4654, 1993.
Bell CC, Finger TE, Russell CJ. Central connections of the posterior lateral
line lobe in mormyrid fish. Exp Brain Res 42: 9 –22, 1981.
Bell CC, Grant K. Corollary discharge effects and sensory processing in the
mormyromast regions of the mormyrid electrosensory lobe. II. Cell types
and corollary discharge plasticity. J Neurophysiol 68: 859 – 875, 1992.
Bell CC, Grant K, Serrier J. Corollary discharge effects and sensory
processing in the mormyrid electrosensory lobe. I. Field potentials and
cellular activity in associated structures. J Neurophysiol 68: 843– 858, 1992.
Bell CC, Han V, Sawtell NB. Cerebellum-like structures and their implications for cerebellar function. Annu Rev Neurosci 31: 1–24, 2008.
Bell CC, Han VZ, Sugawara S, Grant K. Synaptic plasticity in a cerebellumlike structure depends on temporal order. Nature 387: 278 –281, 1997b.
Bell CC, Russell CJ. Effect of electric organ discharge on ampullary receptors
in a mormyrid. Brain Res 145: 85–96, 1978.
Bell CC, Szabo T. Electroreception in mormyrid fish: central anatomy. In:
Electroreception, edited by Bullock TH, Heiligenberg W. New YorkWiley,
1986, p. 375– 421.
Bodznick D, Montgomery JC, Carey M. Adaptive mechanisms in the
elasmobranch hindbrain. J Exp Biol 202: 1357–1364, 1999.
Bol K, Marsat G, Harvey-Girard E, Longtin A, Maler L. Frequency-tuned
cerebellar channels and burst-induced LTD lead to the cancellation of
redundant sensory inputs. J Neurosci 31: 11028 –11038, 2011.
Campbell HR, Meek J, Zhang J, Bell CC. Anatomy of the posterior caudal
lobe of the cerebellum and the eminentia granularis posterior in a mormyrid
fish. J Comp Neurol 502: 714 –735, 2007.
Carey MR, Regehr WG. Noradrenergic control of associative synaptic
plasticity by selective modulation of instructive signals. Neuron 62: 112–
122, 2009.
Cerminara NL, Rawson JA. Evidence that climbing fibers control an intrinsic
spike generator in cerebellar Purkinje cells. J Neurosci 24: 4510 – 4517,
2004.
Colin F, Manil J, Desclin JC. The olivocerebellar system. I. Delayed and
slow inhibitory effects: an overlooked salient feature of cerebellar climbing
fibers. Brain Res 187: 3–27, 1980.
Cowan NJ, Fortune ES. The critical role of locomotion mechanics in
decoding sensory systems. J Neurosci 27: 1123–1128, 2007.
de Ruiter MM, De Zeeuw CI, Hansel C. Voltage-gated sodium channels in
cerebellar Purkinje cells of mormyrid fish. J Neurophysiol 96: 378 –390,
2006.
Dean P, Porrill J, Ekerot CF, Jorntell H. The cerebellar microcircuit as an
adaptive filter: experimental and computational evidence. Nat Rev Neurosci
11: 30 – 43, 2010.
Demer JL, Echelman DA, Robinson DA. Effects of electrical stimulation and
reversible lesions of the olivocerebellar pathway on Purkinje cell activity in
the flocculus of the cat. Brain Res 346: 22–31, 1985.
Ebner TJ, Pasalar S. Cerebellum predicts the future motor state. Cerebellum
7: 583–588, 2008.
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
presumably receive similar EOCD input via parallel fibers
(compare Fig. 3, A and C). Diversity in Purkinje cell EOCD
responses observed in our experimental preparation may reflect
the maintenance of temporally specific plastic changes of
EOCD inputs induced during the fish’s life. A lack of such
diversity in MG cells may reflect the rapid adjustment of the
strength of EOCD inputs to the artificial conditions imposed by
our preparation, in which the EOD motor command has no
sensory consequence because of blockade of the EOD by the
paralytic (Bell 1986a). The stereotyped short-latency excitation
observed in MG cells (Fig. 3C) is believed to be due to a
nonplastic EOCD input from the juxtalobar nucleus to the
basilar dendrites of MG cells (Mohr et al. 2003b).
Finally, it is possible that some additional factor— besides
coactivation of parallel fibers and CFs—is required for plasticity induction in Purkinje cells. Such a requirement would
distinguish Purkinje cell plasticity from that observed in ELL,
where coactivation of parallel fiber inputs with postsynaptic
spikes appears to be the only requirement for plasticity induction and negative image formation. Consistent with such a
possibility, neuromodulatory systems have been shown to
affect induction of CF-induced plasticity of parallel fiber synapses onto Purkinje cells in vitro (Carey and Regehr 2009).
Moreover, a recent study has provided evidence that CF error
signals may be gated, such that they play a role in inducing
Purkinje cell plasticity in the context of some forms of motor
learning but not others (Kimpo et al. 2013).
In summary, our results provide basic information about in
vivo response properties of LCp Purkinje cells that will help
guide future studies of this interesting region of the cerebellum.
Given the key role for motor corollary discharge signals in
hypothesized predictive functions of the cerebellum (Miall and
Wolpert 1996; Wolpert et al. 1998), the prominence and
accessibility of EOCD signals in LCp may provide unique
opportunities for understanding how corollary discharge signals are processed and transformed in cerebellar circuitry.
Moreover, given the striking anatomical parallelism between
ELL and LCp, further studies may also contribute to a general
understanding of how similarities and differences in circuit
architecture relate to function.
SENSORY AND MOTOR RESPONSES OF MORMYRID PURKINJE CELLS
Miall RC, Wolpert DM. Forward models for physiological motor control.
Neural Netw 8: 1–15, 1996.
Mohr C, Roberts PD, Bell CC. Cells of the mormyromast region of the
mormyrid electrosensory lobe. I. Responses to the electric organ corollary
discharge and to electrosensory stimuli. J Neurophysiol 90: 1193–1210,
2003a.
Mohr C, Roberts PD, Bell CC. The mormyromast region of the mormyrid
electrosensory lobe. II. Responses to input from central sources. J Neurophysiol 90: 1211–1223, 2003b.
Montarolo PG, Palestini M, Strata P. The inhibitory effect of the olivocerebellar input on the cerebellar Purkinje cells in the rat. J Physiol 332:
187–202, 1982.
Nieuwenhuys R, Nicholson C. A survey of the general morphology, the fiber
connections, and the possible functional significance of the gigantocerebellum of Mormyrid fishes. In: Neurobiology of Cerebellar Evolution and
Development, edited by Llinas R. Chicago, IL: Am. Med. Assoc., 1969,
107–134.
Post N, von der Emde G. The “novelty response” in an electric fish: response
properties and habituation. Physiol Behav 68: 115–128, 1999.
Roberts PD, Bell CC. Computational consequences of temporally asymmetric
learning rules. II. Sensory image cancellation. J Comput Neurosci 9: 67– 83,
2000.
Rose GJ, Canfield JG. Longitudinal tracking responses of the weakly electric
fish, Sternopygus. J Comp Physiol A 171: 791–798, 1993.
Russell CJ, Bell CC. Neuronal responses to electrosensory input in the
mormyrid valvula cerebelli. J Neurophysiol 41: 1495–1510, 1978.
Savio T, Tempia F. On the Purkinje cell activity increase induced by
suppression of inferior olive activity. Exp Brain Res 57: 456 – 463, 1985.
Sawtell NB. Multimodal integration in granule cells as a basis for associative
plasticity and sensory prediction in a cerebellum-like circuit. Neuron 66:
573–584, 2010.
Sawtell NB, Mohr C, Bell CC. Recurrent feedback in the mormyrid electrosensory system: cells of the preeminential and lateral toral nuclei. J Neurophysiol 93: 2090 –2103, 2005.
Sawtell NB, Williams A. Transformations of electrosensory encoding associated with an adaptive filter. J Neurosci 28: 1598 –1612, 2008.
Sawtell NB, Williams A, Bell CC. Central control of dendritic spikes shapes
the responses of Purkinje-like cells through spike timing-dependent synaptic
plasticity. J Neurosci 27: 1552–1565, 2007.
von der Emde G, Bell C. Nucleus preeminentialis of mormyrid fish, a center
for recurrent electrosensory feedback. I. Electrosensory and corollary discharge responses. J Neurophysiol 76: 1581–1596, 1996.
Wolpert DM, Miall C, Kawato M. Internal models in the cerebellum. Trends
Cogn Sci 2: 338 –347, 1998.
Womack MD, Khodakhah K. Dendritic control of spontaneous bursting in
cerebellar Purkinje cells. J Neurosci 24: 3511–3521, 2004.
Zhang Y, Han VZ. Physiology of morphologically identified cells in the
posterior caudal lobe of the mormyrid cerebellum. J Neurophysiol 98:
1297–1308, 2007.
Zhang Y, Magnus G, Han VZ. Electrophysiological characteristics of cells in
the anterior caudal lobe of the mormyrid cerebellum. Neuroscience 171:
79 –91, 2010.
Zhang Y, Magnus G, Han VZ. Synaptic dynamics and long-term plasticity at
synapses of Purkinje cells onto neighboring Purkinje cells of a mormyrid
fish: a dual cell recording study. Neuroscience 225: 199 –212, 2012.
J Neurophysiol • doi:10.1152/jn.00016.2014 • www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on October 26, 2016
Engelmann J, van den Berg E, Bacelo J, de Ruijters M, Kuwana S,
Sugawara Y, Grant K. Dendritic backpropagation and synaptic plasticity
in the mormyrid electrosensory lobe. J Physiol (Paris) 102: 233–245, 2008.
Fujita M. Adaptive filter model of the cerebellum. Biol Cybern 45: 195–206,
1982.
Grant K, Sugawara S, Gomez L, Han VZ, Bell CC. The mormyrid
electrosensory lobe in vitro: physiology and pharmacology of cells and
circuits. J Neurosci 18: 6009 – 6025, 1998.
Han VZ, Bell CC. Physiology of cells in the central lobes of the mormyrid
cerebellum. J Neurosci 23: 11147–11157, 2003.
Han VZ, Zhang Y, Bell CC, Hansel C. Synaptic plasticity and calcium
signaling in Purkinje cells of the central cerebellar lobes of mormyrid fish.
J Neurosci 27: 13499 –13512, 2007.
Izawa J, Criscimagna-Hemminger SE, Shadmehr R. Cerebellar contributions to reach adaptation and learning sensory consequences of action. J
Neurosci 32: 4230 – 4239, 2012.
Jirenhed DA, Bengtsson F, Hesslow G. Acquisition, extinction, and reacquisition of a cerebellar cortical memory trace. J Neurosci 27: 2493–2502,
2007.
Jirenhed DA, Hesslow G. Learning stimulus intervals—adaptive timing of
conditioned purkinje cell responses. Cerebellum 10: 523–535, 2011.
Kennedy A, Wayne G, Kaifosh P, Alvina K, Abbott LF, Sawtell NB. A
temporal basis for predicting the sensory consequences of motor commands
in an electric fish. Nat Neurosci 17: 416 – 422, 2014.
Kimpo RR, Rinaldi JM, Kim CK, Raymond JL. Gating of error signals in
the cerebellum. Soc Neurosci Abstr 2013: 164.–12, 2013.
Llinas R, Sugimori M. Electrophysiological properties of in vitro Purkinje
cell dendrites in mammalian cerebellar slices. J Physiol 305: 197–213,
1980a.
Llinas R, Sugimori M. Electrophysiological properties of in vitro Purkinje
cell somata in mammalian cerebellar slices. J Physiol 305: 171–195, 1980b.
Marr D. A theory of cerebellar cortex. J Physiol 202: 437– 471, 1969.
Mathews PJ, Lee KH, Peng Z, Houser CR, Otis TS. Effects of climbing
fiber driven inhibition on Purkinje neuron spiking. J Neurosci 32: 17988 –
17997, 2012.
McKay BE, Engbers JD, Mehaffey WH, Gordon GR, Molineux ML, Bains
JS, Turner RW. Climbing fiber discharge regulates cerebellar functions by
controlling the intrinsic characteristics of Purkinje cell output. J Neurophysiol 97: 2590 –2604, 2007.
Medina JF, Garcia KS, Nores WL, Taylor NM, Mauk MD. Timing
mechanisms in the cerebellum: testing predictions of a large-scale computer
simulation. J Neurosci 20: 5516 –5525, 2000.
Medina JF, Mauk MD. Computer simulation of cerebellar information
processing. Nat Neurosci 3, Suppl: 1205–1211, 2000.
Meek J, Yang JY, Han VZ, Bell CC. Morphological analysis of the
mormyrid cerebellum using immunohistochemistry, with emphasis on the
unusual neuronal organization of the valvula. J Comp Neurol 510: 396 – 421,
2008.
Metzen MG, Engelmann J, Bacelo J, Grant K, von der Emde G. Receptive
field properties of neurons in the electrosensory lateral line lobe of the
weakly electric fish, Gnathonemus petersii. J Comp Physiol A Neuroethol
Sens Neural Behav Physiol 194: 1063–1075, 2008.
Miall RC, Weir DJ, Wolpert DM, Stein JF. Is the cerebellum a Smith
predictor? J Mot Behav 25: 203–216, 1993.
339