Articles in PresS. J Neurophysiol (December 24, 2014). doi:10.1152/jn.00587.2014
Acute Testosterone Modulates Voice Perception
1
Testosterone Modulates Preattentive Sensory Processing and Involuntary
2
Attention Switches to Emotional Voices
3
Chenyi Chen1*, Chin-Yau Chen2*, Chih-Yung Yang3, 4*, Chi-Hung Lin3, 4, Yawei Cheng1, 3,
4
5 CA
5
1. Institute of Neuroscience, National Yang-Ming University, Taipei, Taiwan
6
2. Department of Surgery, National Yang-Ming University Hospital, Yilan, Taiwan
7
3. Department of Education and Research, Taipei City Hospital, Taipei, Taiwan
8
4. Institute of Microbiology and Immunology, National Yang-Ming University, Taipei,
9
Taiwan
10
5. Department of Rehabilitation, National Yang-Ming University Hospital, Yilan,
11
Taiwan
12
* Chenyi Chen, Chin-Yau Chen, and Chih-Yung Yang equally contributed to the study.
13
14
Corresponding Author: Prof. Yawei Cheng
15
Institute of Neuroscience
16
National Yang-Ming University
17
155, Sec. 2, St. Linong, Dist. Beitou
18
Taipei, Taiwan 112, R.O.C.
19
Tel: 886-2-28267912
20
Fax: 886-2-28264903
21
Email: [email protected]
22
23
Running Title: Acute Testosterone Modulates Voice Perception
24
Abstract (160 words), text (4055), and 4 figures
1
Copyright © 2014 by the American Physiological Society.
Acute Testosterone Modulates Voice Perception
25
ABSTRACT
26
Testosterone is capable of altering facial threat processing. Voices, similar to faces,
27
convey social information. We hypothesized that administering a single dose of
28
testosterone would change voice perception in humans. In a placebo-controlled,
29
randomly assigned, double-blind crossover design, we administered a single dose of
30
testosterone or placebo to 18 healthy female volunteers and used a passive auditory
31
oddball paradigm. The mismatch negativity (MMN) and P3a in responses to fearfully,
32
happily, and neutrally spoken syllables dada and acoustically matched nonvocal
33
sounds were analyzed, indicating preattentive sensory processing and involuntary
34
attention switches. Results showed that testosterone administration had a trend to
35
shorten the peak latencies of happy MMN and significantly enhanced the amplitudes
36
of happy and fearful P3a, whereas the happy- and fearful-derived nonvocal MMN
37
and P3a remained unaffected. These findings demonstrated acute effect of
38
testosterone on the neural dynamics of voice perception. Administering a single
39
dose of testosterone modulates preattentive sensory processing and involuntary
40
attention switches in response to emotional voices.
41
Keywords: testosterone; voice; mismatch negativity (MMN); P3a
42
43
44
45
46
47
48
2
Acute Testosterone Modulates Voice Perception
49
INTRODUCTION
50
The fear-reducing properties of testosterone have been established and verified
51
using an excessive range of behavioral assessment paradigms in animals (e.g., Aikey
52
et al. 2002; Boissy and Bouissou 1994; Bouissou and Vandenheede 1996; Frye and
53
Seliga 2001) and in humans (e.g., Burris et al. 1992; Pope et al. 2003; van Honk et al.
54
2005; Wang et al. 1996). These findings, in relation to humans, were primarily based
55
on people’s responses to threatening (angry, fearful) faces. Human voices, similar to
56
faces, can convey a wealth of social information (Belin et al. 2004). However, little is
57
known regarding the effect of acute testosterone administration on emotional voice
58
processing.
59
60
Data from animal research indicate that testosterone predominantly affects emotion
61
and motivation by binding to steroid receptive neurons that occupy the amygdaloid-
62
centered nuclei in the limbic system (Schulkin 2003; Wood 1996). Research on
63
humans suggests the crucial role of amygdala in the effects of testosterone on
64
emotional processing (van Wingen et al. 2009). In healthy volunteers, a single
65
administration of testosterone reduced the recognition of angry and fearful facial
66
expressions (van Honk et al. 2005; van Honk et al. 2001). Vigilant attention
67
responses to angry faces were also positively related to testosterone levels (van
68
Honk et al. 1999).
69
70
Mismatch negativity (MMN) and P3a, which are event-related potentials (ERPs), can
71
be elicited by a passive auditory oddball paradigm in which participants engage in a
72
task and ignore the stimuli that are presented in a random series, with one stimulus
3
Acute Testosterone Modulates Voice Perception
73
(standard) occurring more frequently than the other stimulus (deviant). MMN, which
74
is elicited irrespective of the person’s direction of attention (Näätänen et al. 1978), is
75
presumed to reflect the preattentive sensory processing and the function of the N-
76
methyl-D-aspartate (NMDA) receptor (Näätänen et al. 2007). The modeling of
77
generator sources suggests that predominantly frontocentral scalp distribution of
78
MMN is primarily explained by the activity in the bilateral supratemporal cortex
79
(Rinne et al. 1999). P3a, which is the ERP component that occurs after MMN is
80
induced in response to deviants in the unattended stream (Näätänen et al. 2011), is
81
associated with involuntary attention switches caused by sound changes (Escera et
82
al. 2000). Given an increasing visuospatial attentional load, the P3a response
83
becomes smaller in relation to decreased involuntary attention switching (Fan et al.
84
2013; Zhang et al. 2006). MMN and P3a in response to emotional syllables have
85
recently indicated the emotional salience of voice perception at the preattentive
86
level (Chen et al. 2014; Cheng et al. 2012; Fan et al. 2013). Similarly, youths with
87
autism spectrum conditions and conduct disorder symptoms exhibited altered MMN
88
and P3a in response to emotional syllables (Fan and Cheng 2014; Hung et al. 2013).
89
In a previous study, female participants displayed stronger fearful MMN than males
90
did (Hung and Cheng 2014), inferring the sex hormone-mediated processing of
91
emotional voices. However, these findings are only correlational, and do not clarify
92
the causal relation between testosterone and emotional voices.
93
94
To explore the possible causal role of testosterone in the perception of emotional
95
voice and, in particular, to test whether testosterone modulates the neural
96
correlates of fearful voice processing, we investigated the effect of a single
4
Acute Testosterone Modulates Voice Perception
97
administration of testosterone on the presentation of the emotionally spoken
98
syllables “dada” and acoustically matched nonvocal sounds, respectively, in a passive
99
oddball paradigm. Based on a randomized, double-blind, crossover design, we
100
sublingually administered either 0.5 mg of testosterone or a placebo to female
101
volunteers on two separate days. Only women participated, because the quantity
102
and time course for inducing neurophysiological effects in women after a single
103
sublingual administration of 0.5 mg of testosterone have been established, but are
104
unknown for men (Tuiten et al. 2000). We hypothesized that, if testosterone were
105
involved in the preattentive sensory processing of voices, then MMN in response to
106
emotional syllables, but not acoustically matched nonvocal sounds, would be altered
107
after testosterone administration. If involuntary attention switches to voices were
108
related to testosterone, P3a in relation to emotional syllables, rather than
109
acoustically matched nonvocal sounds, would be changed accordingly. Furthermore,
110
this design might enable us to clarify whether the fear-reducing effect of
111
testosterone would be associated with the decreased neural response to fearful
112
voices or the increased neural response to both happy and fearful voices (Archer
113
2006; Bos et al. 2012; Bos et al. 2013; Wingfield et al. 1990). In addition, considering
114
that the right-hand second digit-to-fourth digit (2D: 4D) ratio as a proxy of fetal
115
testosterone is substantiated by evidence derived from animals and humans
116
(Hönekopp et al. 2007; van Honk et al. 2011), we conducted analyses on the
117
correlation of the 2D: 4D ratio with MMN and P3a, respectively. Given that
118
intertribal variability is characteristic of the ERP signal (Spencer 2005), we measured
119
both the amplitude and peak latency of MMN and P3a.
120
5
Acute Testosterone Modulates Voice Perception
121
MATERIALS and METHODS
122
Participants
123
Twenty-one healthy right-handed females, aged between 19 and 27 years (23 ± 1.8
124
years), participated in this study after providing written informed consent. As a
125
result of poor electroencephalogram (EEG) qualities and less than a 2-fold increase in
126
testosterone levels after administration, a total of 18 participants were finally
127
included in the data analysis. They had normal menstruation cycles and were
128
nonsmokers. All the participants had normal hearing; did not have neurological,
129
endocrinal, and psychiatric disorders; and were not taking any medication at the
130
time of testing. This study was approved by the ethics committee at Taipei City
131
Hospital and conducted in accordance with the Declaration of Helsinki. The
132
participants received monetary compensation for their participation.
133
134
Testosterone Administration
135
In the double-blind, crossover, within-subjects design used in this study, the
136
participants received a single dose of 0.5 mg of testosterone on one day and a single
137
dose of placebo 48 h later. The drug samples consisted of 0.5 mg of testosterone
138
undecanoate (Andriol®, N.V. Organon). Testosterone was omitted from the placebo
139
samples, and both testosterone and placebo were administered sublingually. The
140
dosage of 0.5 mg should be sufficient to result in a 10-fold blood testosterone level
141
within 90 min (Tuiten et al. 2000). The interval between two administrations, 48 h,
142
was used to enable exogenous testosterone to metabolize (Slater et al. 2001).
143
144
Auditory Stimuli
6
Acute Testosterone Modulates Voice Perception
145
The stimulus materials consisted of two categories: emotional syllables and
146
acoustically matched nonvocal sounds. Fearfully, happily, and neutrally spoken
147
syllables dada were selected as the stimuli (please refer to Cheng et al. 2012; Fan
148
and Cheng 2014; Fan et al. 2013; Hung et al. 2013; Hung and Cheng 2014 for
149
validation). To create a set of control stimuli that retain acoustical correspondence,
150
i.e., the temporal and spectral features of emotional syllables, we synthesized
151
nonvocal sounds with Praat (Boersma 2001) and MATLAB (MathWorks, Inc., USA).
152
The fundamental frequency (f0) of emotional (fearful, happy, neutral) syllables were
153
extracted to produce the nonvocal sounds using a sine waveform and then
154
multiplied by the syllable envelope. The duration (550 ms) and loudness (minimum:
155
57 dB; maximum: 62 dB; mean: 59 dB) of the stimuli were controlled for the
156
emotional syllables and nonvocal sounds.
157
158
Digit Ratio Measurement
159
The digit ratio was measured using the scanned images of the right hand of each
160
participant, which is a valid method for measuring finger lengths (van Honk et al.
161
2011). When scanning the images, we ensured that details of major creases could be
162
clearly seen. We used Adobe Photoshop to measure the lengths of the second and
163
fourth digits from the ventral proximal crease of the digit to the fingertip. If a band
164
of creases existed at the base of the digit, the measurement was obtained from the
165
most proximal crease.
166
167
Salivary Testosterone Measurement
7
Acute Testosterone Modulates Voice Perception
168
The level of salivary testosterone indicates the free, unbound, or biologically
169
available portion of testosterone in circulation. We used the method that has been
170
reliably applied to measure free testosterone levels in humans (Arregger et al. 2007).
171
Saliva was collected within 15 min before and after the administration of
172
testosterone or placebo. The participants were required to clean their mouth and
173
abstain from drinking and eating for at least one hour before saliva sampling. By
174
chewing a cotton wool swab (Salivette®, Sarstedt) for approximately 60 s, the
175
samples were collected in a Salivette tube. The tubes were subsequently stored at -
176
20°C for further analysis.
177
178
The saliva samples were recovered by centrifuging the Salivette® (combined with the
179
centrifuge vessel) at 1000 x g for 2 min. The clear fluid saliva supernatant was
180
obtained in the conical base of the centrifuge vessel. Based on the manufacturer
181
instructions for an ELISA kit (IBL, Hamburg, Germany), we measured optical density
182
by using a photometer at 450 nm (reference wavelength: 600−650 nm) within 15
183
min after pipetting the Stop Solution for concentration measurements.
184
185
Procedures
186
Figure 1 shows the placebo-controlled, randomly assigned, double-masked, and
187
crossover experimental design. We sublingually administered either 0.5 mg of
188
testosterone or a placebo to female participants on 2 separate days, both of which
189
were conducted 10 days after the menstruation period of each participant had
190
ended. To control the diurnal effect of hormones, we performed all tests in the
191
morning (09:30 AM to 12:30 PM).
8
Acute Testosterone Modulates Voice Perception
192
193
During the passive oddball paradigm for EEG recordings, all participants were
194
required to watch a silent movie with subtitles while task-irrelevant emotional
195
syllables or nonvocal sounds in oddball sequences were presented. The oddball
196
paradigm for emotional syllables employed happily and fearfully spoken syllables
197
dada as deviants and neutrally spoken syllables dada as the standard. The
198
corresponding nonvocal sounds were applied with the same oddball paradigm but
199
were presented in separated blocks. Each stimulus categories comprised two blocks.
200
Each block consisted of 500 trials, of which 80% were neutral syllables or sounds,
201
10% were fearful syllables or sounds, and the other 10% were happy syllables or
202
sounds. The sequences of blocks and stimuli were quasirandomized to avoid
203
successive blocks from identical stimulus categories and successive deviant stimuli.
204
The stimulus-onset-asynchrony was 1200 ms (550 stimulus, 650 idle).
205
206
To control for the potential secondary mood-generated effects of testosterone on
207
voice perception, we used the shortened version of the Profile of Mood States
208
(POMS) (Shacham 1983) for analyzing the participants who received the placebo and
209
testosterone.
210
211
Considering that assumptions regarding the effects of the testosterone hormone can
212
influence the performance of human participants under experimental conditions in
213
which participants believe they have been administered the hormone (Eisenegger et
214
al. 2010), we asked the participants to indicate (by forced choice) in which sessions
215
they believe they received testosterone or placebo after the two administrations.
9
Acute Testosterone Modulates Voice Perception
216
217
Electroencephalogram Apparatus and Recording
218
The EEG was continuously recorded from 32 scalp sites using electrodes mounted in
219
an elastic cap and positioned according to the modified International 10−20 system,
220
with the addition of two mastoid electrodes. The electrode at right mastoid (A2) was
221
used as on-line reference. Eye blinks and vertical eye movements were monitored
222
with electrodes located above and below left eye. The horizontal electro-oculogram
223
(EOG) was recorded from electrodes placed 1.5 cm lateral to the left and right
224
external canthi. A ground electrode was placed on the forehead. Electrode/skin
225
impedance was kept < 5 k. Channels were re-referenced off-line to the average of
226
left and right mastoid recordings [(A1 + A2)/2]. Signals were sampled at 500 Hz,
227
band-pass filtered (0.1−100 Hz), and epoched over an analysis time of 600 ms,
228
including pre-stimulus 100 ms used for baseline correction. An automatic artifact
229
rejection system excluded from the average all trials containing transients exceeding
230
70V at recording electrodes and exceeding 100 V at the horizontal EOG
231
channel. Furthermore, the quality of ERP traces was ensured by careful visual
232
inspection in every subject and trial and by applying appropriate digital, zero-phase
233
shift band-pass filtered (0.1–50 Hz, 24 dB/octave). The first ten epochs of each
234
sequence were omitted from the averaging in order to exclude unexpected large
235
responses elicited by the initiation of the sequences.
236
237
The paradigm was edited by MATLAB (MathWorks, Inc., USA). Each event in the
238
paradigm was associated with a digital code that was sent to the continuous EEG,
10
Acute Testosterone Modulates Voice Perception
239
thereby enabling the off-line segmentation and the average of selected EEG periods
240
to be obtained for analysis. The ERPs were processed and analyzed using Neuroscan
241
4.3 (Compumedics Ltd., Australia).
242
243
Statistical Analysis
244
The amplitudes of MMN and P3a were analyzed as an average within a 50-ms time
245
window surrounding the peak latency at the electrode sites F3, Fz, F4, C3, Cz, and C4.
246
The MMN peak was defined as the largest negativity in the subtraction between the
247
deviant and standard ERPs, during a period of 150 to 250 ms after sound onset. Only
248
the standards presented before the deviants were included in the analysis. The P3a
249
peak was defined as the highest positivity during a period of 300 to 500 ms.
250
251
Statistical analyses on MMN and P3a were conducted using four-way ANOVA
252
comprising the factors of the administration (testosterone or placebo), category
253
(emotional syllables or nonvocal sounds), affect (fearful or happy), and electrode (F3,
254
Fz, F4, C3, Cz, or C4). The dependent variables were the mean amplitudes and peak
255
latencies of the MMN and P3a components at the selected electrode sites. Degrees
256
of freedom were corrected using the Greenhouse-Geisser method. The post hoc
257
comparisons were only conducted when preceded by significant main effects.
258
259
RESULTS
260
Mood Measurement
261
The shortened version of POMS was used to indicate the possible effects of
262
testosterone on tension-anxiety, depression, anger, fatigue, and vigor. The Wilcoxon
11
Acute Testosterone Modulates Voice Perception
263
signed-rank test revealed no significant differences in mood between those who
264
received the placebo and testosterone (p > 0.11). Given that testosterone did not
265
affect mood, the observed effects of testosterone on emotional voice processing
266
cannot be attributed to secondary mood-generated response biases.
267
268
Control of Belief Effects and Subjective Biases
269
Performance was at the chance level (N = 21, binomial p = 0.78), which confirmed
270
that the participants were unaware of what they had received.
271
272
Salivary Testosterone
273
The
274
administration, but 105% after placebo [t(17) = 6.52, p < .001]. The testosterone and
275
placebo concentrations were 1908.64 ± 481.80 pg/mL and 265.62 ± 133.08 pg/mL,
276
respectively.
salivary
testosterone
level
increased
to
1061%
after
testosterone
277
278
MMN elicited by Emotional Syllables and Nonvocal Sounds
279
The preattentive discrimination of emotional voices was studied using MMN, and
280
was determined by subtracting the neutral ERP from fearful and happy ERPs (Figure
281
2). The repeated measures ANOVA model was used to analyze MMN amplitudes and
282
revealed that the main effects of category [emotional syllables vs. nonvocal sounds:
283
F(1, 17) = 11.35, p = .004, η2 = 0.40], affect [fearful vs. happy: F(1, 17) = 62.44, p <
284
.001, η2 = 0.79], and electrode [F(5, 85) = 7.69, p = .001, η2 = 0.31] exerted
285
significance (Figure 3). Emotional syllables elicited stronger MMN than did nonvocal
286
sounds. Fearful MMN had stronger amplitudes than did happy MMN. Fz and Cz had
12
Acute Testosterone Modulates Voice Perception
287
more negative deflections than did other electrodes (Fz vs. F3: Bonferroni-corrected
288
p = .001; Fz vs. C3: p = .004; Fz vs. C4: p = .024; Cz vs. C3; p = .001; Cz vs. C4: p =
289
.005). However, neither the administration [testosterone vs. placebo: F(1, 17) = 1.82,
290
p = .20, η2 = 0.10] nor related interactions [administration x category: F(1, 17) = 0.44,
291
p = .52, η2 = 0.03; administration x affect: F(1, 17) = 0.102, p = .75, η2 = 0.01;
292
administration x category x affect: F(1, 17) = 0.16, p = .69, η2 = 0.09] reached
293
statistical significance.
294
295
The repeated measures ANOVA conducted on MMN peak latencies revealed that the
296
affect (fearful vs. happy) produced a main effect [F(1, 17) = 12.34, p = .003, η2 =
297
0.42], but the administration [F(1, 17) = 0.24, p = .63, η2 = 0.01], category [emotional
298
syllables vs. nonvocal sounds: F(1, 17) = 4.11, p = .059, η2 = 0.20], and electrode [F(5,
299
85) = 1.85, p = .16, η2 = 0.10] exerted no effects. Fearful MMN had longer latencies
300
than did happy MMN. In addition, an interaction occurred between administration
301
and category [F(1, 17) = 5.85, p = .027, η2 = 0.26], but no interaction occurred
302
between administration and affect [F(1, 17) = 0.22, p = .64, η2 = 0.01]. Emotional
303
MMN relative to nonvocal MMN had longer latencies after placebo (p = .007),
304
whereas this effect was diminished after testosterone administration (p = .90).
305
Particularly, the interaction among administration, category, and affect had a
306
marginal significance with medium to large effect size [F(1, 17) = 4.11, p = .059, η2 =
307
0.20]. Post hoc analysis indicated that happy MMN exerted an interaction between
308
administration and category [F(1, 17) = 5.36, p = .033, η2 = 0.24], but fearful MMN
309
did not [F(1, 17) = 0.21, p = .65, η2 = 0.01]. The administration effect of happy MMN
310
showed a trend that approached significance [t(17) = 1.828, p = .085, d = 0.43],
13
Acute Testosterone Modulates Voice Perception
311
indicating that testosterone relative to placebo administration tended to shorten
312
happy MMN latencies. Instead, fearful MMN failed to achieve a customary level of
313
statistical significance [t(17) = -1.451, p = .165, d = 0.34] (Figure 4).
314
315
P3a elicited by Emotional Syllables and Nonvocal Sounds
316
The repeated measures ANOVAs of P3a amplitudes revealed main effects produced
317
by the administration [testosterone vs. placebo: F(1, 17) = 6.61, p =.02, η2 = 0.28],
318
category [emotional syllables vs. nonvocal sounds: F(1, 17) = 4.79, p = .043, η2 =
319
0.22], affect [fearful vs. happy: F(1, 17) = 76.29, p < .001, η2 = 0.82], and electrode
320
[F(5, 85) = 27.85, p < .001, η2 = 0.62] (Figure 3). P3a significantly differed between
321
the administration of testosterone and placebo. Emotional syllables elicited stronger
322
P3a than did nonvocal sounds. Fearful P3a exhibited stronger amplitudes than did
323
happy P3a. The midline electrodes had larger P3a than did the lateral electrodes (Fz
324
vs. F3: Bonferroni-corrected p < .001; Fz vs. F4: p < .001; Cz vs. C3: p < .001; Cz vs. C4:
325
p < .001). Moreover, a significant interaction occurred between administration and
326
category [F(1, 17) = 11.10, p = .004, η2 = 0.40], but none occurred between
327
administration and affect [F(1, 17) = 0.87, p = .36, η2 = 0.05] nor among
328
administration, category, and affect [F(1, 17) = 0.14, p = .72, η2 = 0.01]. Follow-up
329
analysis indicated that testosterone relative to placebo administration achieved
330
significance only for emotional P3a (p = .002) rather than nonvocal P3a (p = .32).
331
Testosterone administration increased P3a in response to happy and fearful
332
syllables, but not to acoustically matched nonvocal sounds.
333
14
Acute Testosterone Modulates Voice Perception
334
The ANOVA of P3a peak latencies did not reveal any effect caused by the
335
administration [F(1,17) = 0.47, p = .50, η2 = 0.03], category [F(1, 17) = 2.27, p = .15, η2
336
= 0.12], affect [F(1, 17) = 0.29, p = .60, η2 = 0.02], or electrode [F(5, 85) = 0.75, p =
337
.52, η2 = 0.04] (Figure 4). None of the administration-related interactions reached
338
significance [administration x category: F(1, 17) = 0.82, p = 0.38, η2 = 0.05;
339
administration x affect: F(1, 17) = 0.34, p = .57, η2 = 0.02; administration x category x
340
affect: F(1, 17) = 0.09, p = .77, η2 = 0.01].
341
342
Correlations Between 2D: 4D Ratio and MMN Amplitudes
343
Table 1 lists that neither emotional nor nonvocal MMN and P3a were correlated with
344
the 2D: 4D ratio, irrespective of whether testosterone or placebo was administered
345
(all p > .05).
346
347
DISCUSSION
348
We investigated the causal role of testosterone in the neural correlates of emotional
349
voices. Healthy young women received a single dose of testosterone in a
350
randomized, placebo-controlled, crossover manner, and MMN and P3a elicited by
351
emotional (happy, fearful) syllables and acoustically matched nonvocal sounds were
352
recorded in a passive auditory oddball paradigm. The results indicated that a single
353
dose of testosterone administration tended to accelerate the peak latencies of
354
happy MMN and significantly enhanced the amplitudes of happy and fearful P3a.
355
The 2D: 4D ratio, a proxy of fetal testosterone, was not correlated with emotional
356
MMN and P3a. These findings first demonstrated acute testosterone effect on
357
emotional
voice
processing.
Testosterone
administration
could
modulate
15
Acute Testosterone Modulates Voice Perception
358
preattentive sensory discrimination as well as involuntary attention switches to
359
emotional voices in humans.
360
361
Irrespective of amplitudes and latencies, a single administration of testosterone did
362
not change fearful MMN. Sex differences in fearful MMN identified in young adults
363
(Hung and Cheng 2014) are not likely attributed to acute testosterone effect.
364
Theoretically, three possible reasons can explain this phenomenon. First, the cortical
365
generator of MMN likely originates from the primary auditory cortex or its vicinity
366
(Alho 1995). The insula, located near the primary auditory cortex, is capable of
367
integrating multisensory modalities (Rodgers et al. 2008) and is involved in
368
processing emotional voices (Belin et al. 2004; Chen et al. 2014; Morris et al. 1999;
369
Sander and Scheich 2001). However, no evidence of androgen receptors in the insula
370
of humans is currently available (Bancroft 2005). If the insula underpins emotional
371
voice processing at the preattentive level, acute testosterone administration would
372
not modulate fearful MMN. Second, physiological levels of testosterone may be
373
necessary for obtaining a generalized pattern of interactive brain activity, much of
374
which is independent of the direct testosterone effect. Currently, only a limited
375
overlap exists between localized testosterone effects in the brain based on a
376
comparison androgen receptors and neuroimaging studies (Peper et al. 2011). This
377
may partially reflect the methodological limitations of ERPs; certain relevant areas of
378
the brain, particularly the amygdala, are too deep to be recorded using this method.
379
These two techniques are used to examine extremely diverse markers of brain
380
activity − steroid receptors and postsynaptic potentials. Third, the testosterone
381
effect on MMN does not appear without long-term exposure. NMDA receptor
16
Acute Testosterone Modulates Voice Perception
382
antagonists and agonists affect MMN generation (e.g., Javitt et al. 1996; Leung et al.
383
2008; Umbricht et al. 2000). Testosterone modulates NMDA receptor-mediated
384
mRNA levels and postsynaptic currents (Brann et al. 1993; White et al. 1999).
385
Testosterone acts through NMDA receptors to affect the motor neuron response to
386
glutamate and blockage of NMDA receptors decreases testosterone levels (Hsu et al.
387
2000). Accordingly, acute testosterone effect on MMN will not be evident without
388
following these cascade steps.
389
390
A single dose of testosterone administration enhances P3a responses to emotional
391
(happy, fearful) syllables in a passive auditory oddball paradigm. Testosterone plays
392
a crucial role in affective behavior, which is mediated by various brain regions within
393
the emotional circuitry, including the prefrontal cortex and amygdala (van Wingen et
394
al. 2010; Volman et al. 2011). Endogenous testosterone levels were related to
395
attentional orienting to angry faces (van Honk et al. 1999; Wirth and Schultheiss
396
2007) and smiling faces (Cashdan 1995; Dabbs 1997). Exogenous testosterone
397
moderated attentional responses to threatening stimuli (Lacreuse et al. 2010; van
398
Honk et al. 2005). Testosterone might particularly increase attentional biases to
399
negative stimuli during early exposure to the stimuli accompanied by acute
400
treatment, or when the stimuli are highly aroused by chronic treatment (King et al.
401
2012). P3a reflects involuntary attentional switches caused by novel stimuli
402
(Harmony et al. 2000). Lesion studies have indicated that the prefrontal cortex might
403
contribute to P3a generation (Knight 1984). This study demonstrated that exogenous
404
testosterone enhanced involuntary attention switches to emotional voices, as
405
indicated by heightened responses to both happy and fearful P3a. This finding
17
Acute Testosterone Modulates Voice Perception
406
concurs with the hypothesis that the function of testosterone could facilitate any
407
form of adaptive goal-directed behavior, such as the sensitivity to socially relevant
408
signals (Archer 2006; Bos et al. 2012; Wingfield et al. 1990). Interestingly, one recent
409
fMRI study showed that testosterone reduced fear by increasing the amygdala
410
activity to fearful and happy faces (Bos et al. 2013). We thus ascribed the fear-
411
reducing properties of testosterone to strengthening, rather than attenuating, the
412
response to emotional faces and voices. These findings might lend support acute
413
testosterone effect to be associated with more elaborate processing of social signals.
414
Incorporating with previous finding that anterior insular cortex activity responded to
415
MMN elicited by hearing disgust (Chen et al. 2014), we supposed that augmented
416
P3a to fearful and happy voices in the present study could be in parallel with
417
enhanced amygdala reactivity to fearful and happy faces (Bos et al. 2013). The
418
source localization of emotional MMN and P3a should be dissociated.
419
420
Particularly, the salivary testosterone level increased to 1061% after testosterone
421
administration with large between-subject variability (385% to 2789%). It could be
422
attributed to the weak serum-saliva correlation for testosterone in our female
423
subjects. Only 1 to 10% of testosterone is in its unbound or biologically active form in
424
blood. The remaining testosterone is bound to serum proteins whereas the majority
425
of testosterone in saliva is not protein-bound. The salivary testosterone levels are
426
modestly correlated with serum levels for males, but not necessarily for females
427
(Shirtcliff et al. 2002). Here, we administered sublingual testosterone in females,
428
whose psychological and physiological effects had been testified by an amount of
429
studies (e.g., Slater et al. 2001; Tuiten et al. 2000; van Honk et al. 2001).
18
Acute Testosterone Modulates Voice Perception
430
431
Taken together, these data first provide the direct evidence for examining acute
432
testosterone effect on the neural dynamics of emotional voice processing in
433
humans. Based on a passive auditory oddball paradigm, testosterone administration
434
tended to advance the peak latencies of MMN in response to happy syllables, as well
435
as significantly increased the amplitudes of P3a to happy and fearful syllables.
436
Testosterone modulated preattentive sensory discrimination and involuntary
437
attention switches to emotional voice processing. The findings suggest that acute
438
testosterone effect is unlikely to cause the emergence of sex differences in fearful
439
MMN during adulthood (Cheng et al. 2012; Hung and Cheng 2014). Considering the
440
role of testosterone in attention control (e.g., Burris et al. 1992; Pope et al. 2003;
441
van Honk et al. 2005; Wang et al. 1996), we argued that testosterone might reduce
442
fear by facilitating preattentive sensory processing of positive social cues as well as
443
elaborate attentional processing of both positive and negative social signals.
444
445
ACKNOWLEDGEMENTS
446
We thank Hou-Ling Lin for assisting with collecting the data.
447
448
GRANTS
449
The study was funded by the Ministry of Science and Technology (MOST 103-2401-
450
H-010-003-MY3), National Yang-Ming University Hospital (RD2014-003), Health
451
Department of Taipei City Government (10301-62-009), and Ministry of Education
452
(Aim for the Top University Plan).
453
19
Acute Testosterone Modulates Voice Perception
454
DISCLOSURES
455
The authors all declare no conflicts of interest, financial, or otherwise.
456
457
REFERENCES
458
Aikey JL, Nyby JG, Anmuth DM, James PJ. Testosterone rapidly reduces anxiety in
459
460
461
462
463
male house mice (Mus musculus). Horm Behav 42: 448-460, 2002.
Alho K. Cerebral generators of mismatch negativity (MMN) and its magnetic
counterpart (MMNm) elicited by sound changes. Ear Hear 16: 38-51, 1995.
Archer J. Testosterone and human aggression: an evaluation of the challenge
hypothesis. Neurosci Biobehav Rev 30: 319-345, 2006.
464
Arregger AL, Contreras LN, Tumilasci OR, Aquilano DR, Cardoso EM. Salivary
465
testosterone: a reliable approach to the diagnosis of male hypogonadism.
466
Clin Endocrinol (Oxf) 67: 656-662, 2007.
467
Bancroft J. The endocrinology of sexual arousal. J Endocrinol 186: 411-427, 2005.
468
Belin P, Fecteau S, Bedard C. Thinking the voice: neural correlates of voice
469
470
471
perception. Trends Cogn Sci 8: 129-135, 2004.
Boersma P. Praat, a system for doing phonetics by computer. Glot International 5:
341–345, 2001.
472
Boissy A, Bouissou MF. Effects of androgen treatment on behavioral and
473
physiological responses of heifers to fear-eliciting situations. Horm Behav 28:
474
66-83, 1994.
475
Bos PA, Panksepp J, Bluthe RM, van Honk J. Acute effects of steroid hormones and
476
neuropeptides on human social-emotional behavior: a review of single
477
administration studies. Front Neuroendocrinol 33: 17-35, 2012.
20
Acute Testosterone Modulates Voice Perception
478
Bos PA, van Honk J, Ramsey NF, Stein DJ, Hermans EJ. Testosterone administration
479
in women increases amygdala responses to fearful and happy faces.
480
Psychoneuroendocrinology 38: 808-817, 2013.
481
482
Bouissou MF, Vandenheede M. Long-term effects of androgen treatment on fear
reactions in ewes. Horm Behav 30: 93-99, 1996.
483
Brann DW, Zamorano PL, Chorich LP, Mahesh VB. Steroid hormone effects on
484
NMDA receptor binding and NMDA receptor mRNA levels in the
485
hypothalamus and cerebral cortex of the adult rat. Neuroendocrinology 58:
486
666-672, 1993.
487
Burris AS, Banks SM, Carter CS, Davidson JM, Sherins RJ. A long-term, prospective
488
study of the physiologic and behavioral effects of hormone replacement in
489
untreated hypogonadal men. J Androl 13: 297-304, 1992.
490
Cashdan E. Hormones, sex, and status in women. Horm Behav 29: 354-366, 1995.
491
Chen C, Lee SH, Cheng Y. Anterior insular cortex activity to emotional salience of
492
voices in a passive oddball paradigm. Front Hum Neurosci 8: 743, 2014.
493
Cheng Y, Lee SY, Chen HY, Wang PY, Decety J. Voice and emotion processing in the
494
495
496
human neonatal brain. J Cogn Neurosci 24: 1411-1419, 2012.
Dabbs JM, Jr. Testosterone, smiling, and facial appearance. J Noverbal Behav 21: 4555, 1997.
497
Eisenegger C, Naef M, Snozzi R, Heinrichs M, Fehr E. Prejudice and truth about the
498
effect of testosterone on human bargaining behaviour. Nature 463: 356-359,
499
2010.
21
Acute Testosterone Modulates Voice Perception
500
Escera C, Alho K, Schröger E, Winkler I. Involuntary attention and distractibility as
501
evaluated with event-related brain potentials. Audiol Neurootol 5: 151-166,
502
2000.
503
Fan YT, Cheng Y. Atypical mismatch negativity in response to emotional voices in
504
people with autism spectrum conditions. PLoS One 9: e102471, 2014.
505
Fan YT, Hsu YY, Cheng Y. Sex matters: n-back modulates emotional mismatch
506
507
508
negativity. Neuroreport 24: 457-463, 2013.
Frye CA, Seliga AM. Testosterone increases analgesia, anxiolysis, and cognitive
performance of male rats. Cogn Affect Behav Neurosci 1: 371-381, 2001.
509
Harmony T, Bernal J, Fernandez T, Silva-Pereyra J, Fernandez-Bouzas A, Marosi E,
510
Rodriguez M, Reyes A. Primary task demands modulate P3a amplitude. Brain
511
Res Cogn Brain Res 9: 53-60, 2000.
512
Hönekopp J, Bartholdt L, Beier L, Liebert A. Second to fourth digit length ratio
513
(2D:4D) and adult sex hormone levels: new data and a meta-analytic review.
514
Psychoneuroendocrinology 32: 313-321, 2007.
515
Hsu C, Hsieh YL, Yang RC, Hsu HK. Blockage of N-methyl-D-aspartate receptors
516
decreases testosterone levels and enhances postnatal neuronal apoptosis in
517
the preoptic area of male rats. Neuroendocrinology 71: 301-307, 2000.
518
Hung AY, Ahveninen J, Cheng Y. Atypical mismatch negativity to distressful voices
519
associated with conduct disorder symptoms. J Child Psychol Psychiatry 54:
520
1016-1027, 2013.
521
522
Hung AY, Cheng Y. Sex differences in preattentive perception of emotional voices
and acoustic attributes. Neuroreport 25: 464-469, 2014.
22
Acute Testosterone Modulates Voice Perception
523
Javitt DC, Steinschneider M, Schroeder CE, Arezzo JC. Role of cortical N-methyl-D-
524
aspartate receptors in auditory sensory memory and mismatch negativity
525
generation: implications for schizophrenia. Proc Natl Acad Sci USA 93: 11962-
526
11967, 1996.
527
King HM, Kurdziel LB, Meyer JS, Lacreuse A. Effects of testosterone on attention and
528
memory
529
Psychoneuroendocrinology 37: 396-409, 2012.
530
531
for
emotional
stimuli
in
male
rhesus
monkeys.
Knight RT. Decreased response to novel stimuli after prefrontal lesions in man.
Electroencephalogr Clin Neurophysiol 59: 9-20, 1984.
532
Lacreuse A, King HM, Kurdziel LB, Partan SR, Caldwell KM, Chiavetta MR, Millette
533
MM, Meyer JS, Grow DR. Testosterone may increase selective attention to
534
threat in young male macaques. Horm Behav 58: 854-863, 2010.
535
Leung S, Croft RJ, O'Neill BV, Nathan PJ. Acute high-dose glycine attenuates
536
mismatch negativity (MMN) in healthy human controls. Psychopharmacology
537
(Berl) 196: 451-460, 2008.
538
539
540
541
Morris JS, Scott SK, Dolan RJ. Saying it with feeling: neural responses to emotional
vocalizations. Neuropsychologia 37: 1155-1163, 1999.
Näätänen R, Gaillard AW, Mantysalo S. Early selective-attention effect on evoked
potential reinterpreted. Acta Psychol (Amst) 42: 313-329, 1978.
542
Näätänen R, Kujala T, Winkler I. Auditory processing that leads to conscious
543
perception: a unique window to central auditory processing opened by the
544
mismatch negativity and related responses. Psychophysiology 48: 4-22, 2011.
23
Acute Testosterone Modulates Voice Perception
545
Näätänen R, Paavilainen P, Rinne T, Alho K. The mismatch negativity (MMN) in basic
546
research of central auditory processing: a review. Clin Neurophysiol 118:
547
2544-2590, 2007.
548
Peper JS, van den Heuvel MP, Mandl RCW, Hulshoff Pol HE, van Honk J. Sex steroids
549
and connectivity in the human brain: a review of neuroimaging studies.
550
Psychoneuroendocrinology 36: 1101-1113, 2011.
551
Pope HG, Jr., Cohane GH, Kanayama G, Siegel AJ, Hudson JI. Testosterone gel
552
supplementation for men with refractory depression: a randomized, placebo-
553
controlled trial. Am J Psychiatry 160: 105-111, 2003.
554
Rinne T, Gratton G, Fabiani M, Cowan N, Maclin E, Stinard A, Sinkkonen J, Alho K,
555
Näätänen R. Scalp-recorded optical signals make sound processing in the
556
auditory cortex visible? Neuroimage 10: 620-624, 1999.
557
558
Rodgers KM, Benison AM, Klein A, Barth DS. Auditory, somatosensory, and
multisensory insular cortex in the rat. Cereb Cortex 18: 2941-2951, 2008.
559
Sander K, Scheich H. Auditory perception of laughing and crying activates human
560
amygdala regardless of attentional state. Brain Res Cogn Brain Res 12: 181-
561
198, 2001.
562
563
564
565
566
567
Schulkin J. Rethinking Homeostasis: Allostatic Regulation in Physiology and
Pathophysiology. Cambridge, MA: MIT Press, 2003.
Shacham S. A shortened version of the Profile of Mood States. J Pers Assess 47: 305306, 1983.
Shirtcliff EA, Granger DA, Likos A. Gender differences in the validity of testosterone
measured in saliva by immunoassay. Horm Behav 42: 62-69, 2002.
24
Acute Testosterone Modulates Voice Perception
568
Slater CC, Souter I, Zhang C, Guan C, Stanczyk FZ, Mishell DR. Pharmacokinetics of
569
testosterone after percutaneous gel or buccal administration. Fertil Steril 76:
570
32-37, 2001.
571
Spencer KM. Averaging, Detection, and Classification of Single-Trial ERPs. In: Event-
572
related potentials: a methods handbook, edited by Handy TC. Cambridge,
573
MA: MIT Press, 2005, p. 57-82.
574
Tuiten A, Van Honk J, Koppeschaar H, Bernaards C, Thijssen J, Verbaten R. Time
575
course of effects of testosterone administration on sexual arousal in women.
576
Arch Gen Psychiatry 57: 149-153; discussion 155-146, 2000.
577
Umbricht D, Schmid L, Koller R, Vollenweider FX, Hell D, Javitt DC. Ketamine-
578
induced deficits in auditory and visual context-dependent processing in
579
healthy volunteers: implications for models of cognitive deficits in
580
schizophrenia. Arch Gen Psychiatry 57: 1139-1147, 2000.
581
van Honk J, Peper JS, Schutter DJ. Testosterone reduces unconscious fear but not
582
consciously experienced anxiety: implications for the disorders of fear and
583
anxiety. Biol Psychiatry 58: 218-225, 2005.
584
van Honk J, Schutter DJ, Bos PA, Kruijt AW, Lentjes EG, Baron-Cohen S.
585
Testosterone administration impairs cognitive empathy in women depending
586
on second-to-fourth digit ratio. Proc Natl Acad Sci USA 108: 3448-3452, 2011.
587
van Honk J, Tuiten A, Hermans E, Putman P, Koppeschaar H, Thijssen J, Verbaten R,
588
van Doornen L. A single administration of testosterone induces cardiac
589
accelerative responses to angry faces in healthy young women. Behav
590
Neurosci 115: 238-242, 2001.
25
Acute Testosterone Modulates Voice Perception
591
van Honk J, Tuiten A, Verbaten R, van den Hout M, Koppeschaar H, Thijssen J, de
592
Haan E. Correlations among salivary testosterone, mood, and selective
593
attention to threat in humans. Horm Behav 36: 17-24, 1999.
594
van Wingen G, Mattern C, Verkes RJ, Buitelaar J, Fernandez G. Testosterone
595
reduces amygdala-orbitofrontal cortex coupling. Psychoneuroendocrinology
596
35: 105-113, 2010.
597
van Wingen GA, Zylicz SA, Pieters S, Mattern C, Verkes RJ, Buitelaar JK, Fernandez
598
G. Testosterone increases amygdala reactivity in middle-aged women to a
599
young adulthood level. Neuropsychopharmacology 34: 539-547, 2009.
600
Volman I, Toni I, Verhagen L, Roelofs K. Endogenous testosterone modulates
601
prefrontal-amygdala connectivity during social emotional behavior. Cereb
602
Cortex 21: 2282-2290, 2011.
603
Wang C, Eyre DR, Clark R, Kleinberg D, Newman C, Iranmanesh A, Veldhuis J,
604
Dudley RE, Berman N, Davidson T, Barstow TJ, Sinow R, Alexander G,
605
Swerdloff RS. Sublingual testosterone replacement improves muscle mass
606
and strength, decreases bone resorption, and increases bone formation
607
markers in hypogonadal men--a clinical research center study. J Clin
608
Endocrinol Metab 81: 3654-3662, 1996.
609
610
White SA, Livingston FS, Mooney R. Androgens modulate NMDA receptor-mediated
EPSCs in the zebra finch song system. J Neurophysiol 82: 2221-2234, 1999.
611
Wingfield JC, Hegner RE, Dufty AM, Ball GF. The “challenge hypothesis”: theoretical
612
implications for patterns of testosterone secretion, mating systems, and
613
breeding strategies. American Naturalist 136: 829-846, 1990.
26
Acute Testosterone Modulates Voice Perception
614
615
616
617
Wirth MM, Schultheiss OC. Basal testosterone moderates responses to anger faces
in humans. Physiol Behav 90: 496-505, 2007.
Wood RI. Functions of the steroid-responsive neural network in the control of male
hamster sexual behavior. Trends Endocrinol Metab 7: 338-344, 1996.
618
Zhang P, Chen X, Yuan P, Zhang D, He S. The effect of visuospatial attentional load
619
on the processing of irrelevant acoustic distractors. Neuroimage 33: 715-724,
620
2006.
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
27
Acute Testosterone Modulates Voice Perception
638
TABLE LEGENDS
639
TABLE 1: Correlations between the 2D: 4D ratio and MMN / P3a amplitudes at Fz (N
640
= 18).
Administration
Testosterone
Placebo
2D: 4D ratio
r
p
Emotional
Happy MMN/P3a
-0.18/-0.13
Fearful MMN/P3a
-0.13/ 0.02
Nonvocal
Happy-derived MMN/P3a -0.15/-0.23
Fearful-derived MMN/P3a -0.11/ 0.13
Emotional
Happy MMN/P3a
-0.22/-0.18
Fearful MMN / P3a
-0.16/ 0.34
Nonvocal
Happy-derived MMN/P3a 0.09/ 0.16
Fearful-derived MMN/P3a -0.02/ 0.18
0.48/0.62
0.62/0.94
0.55/0.35
0.66/0.60
0.38/0.46
0.53/0.16
0.72/0.52
0.95/0.48
641
642
643
FIGURE LEGENDS
644
Figure 1: Experimental procedures.
645
The design is placebo-controlled, randomly assigned, double-masked, and crossover.
646
We sublingually administered either testosterone (T) or placebo (P) to participants
647
on two separate days. Saliva was collected before and after the administration.
648
Participant needed to fill in the Profile of Mood States (POMS). During EEG
649
recordings, participants were required to watch a silent movie with subtitles while
650
task-irrelevant emotional syllables or nonvocal sounds in oddball sequences were
651
presented. The digit ratio was measured using the scanned images of the right hand.
652
Finally, participants had to indicate (guessing) in which sessions they believe they
653
received testosterone or placebo.
654
28
Acute Testosterone Modulates Voice Perception
655
Figure 2: Grand average standard and deviant ERP waveforms in response to
656
emotional syllables and nonvocal sounds after administration of testosterone and
657
placebo.
658
659
Figure 3: Amplitudes of MMN and P3a to emotional syllables and nonvocal sounds
660
after administration of testosterone and placebo.
661
For MMN, neither the administration nor related interactions reached statistical
662
significance. As for P3a, the administration produced a main effect and the
663
interaction between the administration and category reached significance. Follow-up
664
analysis indicated that testosterone relative to placebo administration significantly
665
enhanced fearful and happy P3a (*, p = .002), rather than fearful- and happy-derived
666
nonvocal P3a (p = .32).
667
668
Figure 4: Peak latencies of MMN and P3a to emotional syllables and nonvocal
669
sounds after administration of testosterone and placebo.
670
MMN had a marginal interaction among administration, category, and affect with
671
medium to large effect size (p = .059, η2 = 0.20). Follow-up analysis indicated that
672
testosterone relative to placebo administration tended to shorten happy MMN
673
latencies (+, p < .05, one-tailed), but not to fearful MMN (p = .16). As for P3a, neither
674
the administration nor related interactions reached any statistical significance.
29