Author’s Accepted Manuscript

Human centromedian-parafascicular complex

signals sensory cues for goal-oriented behavior
selection

Inga M. Schepers, Anne-Kathrin Beck, Susann

Bräuer, Kerstin Schwabe, Mahmoud Abdallat,
Pascale Sandmann, Reinhard Dengler, Jochem W.
Rieger, Joachim K. Krauss www.elsevier.com

PII: S1053-8119(17)30227-6
DOI: http://dx.doi.org/10.1016/j.neuroimage.2017.03.019
Reference: YNIMG13894
To appear in: NeuroImage
Received date: 20 December 2016
Revised date: 28 February 2017
Accepted date: 9 March 2017
Cite this article as: Inga M. Schepers, Anne-Kathrin Beck, Susann Bräuer,
Kerstin Schwabe, Mahmoud Abdallat, Pascale Sandmann, Reinhard Dengler,
Jochem W. Rieger and Joachim K. Krauss, Human centromedian-parafascicular
complex signals sensory cues for goal-oriented behavior selection, NeuroImage,
http://dx.doi.org/10.1016/j.neuroimage.2017.03.019
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
Human centromedian-parafascicular complex signals sensory cues for goal-
oriented behavior selection

Inga M. Schepersa,e*1, Anne-Kathrin Beckb,e1, Susann Bräuera,e, Kerstin Schwabeb,e, Mahmoud

Abdallatb, Pascale Sandmannc, Reinhard Denglerd,e, Jochem W. Riegera,e2, Joachim K. Kraussb,e2

a
Department of Psychology, Oldenburg University, Germany
b
Department of Neurosurgery, Hannover Medical School, Germany
c
Department of Otorhinolaryngology, University of Cologne, Cologne, Germany
d
Department of Neurology, Hannover Medical School, Hanover, Germany
e
Cluster of Excellence Hearing4All

E-mail: inga.maren.schepers@uni-oldenburg.de

*Correspondence should be sent to. Inga M. Schepers. Department of Psychology, Faculty VI,
Oldenburg University, 26129 Oldenburg, Germany

Abstract
Experimental research has shown that the centromedian-parafascicular complex (CM-Pf) of the
intralaminar thalamus is activated in attentional orienting and processing of behaviorally relevant
stimuli. These observations resulted in the hypothesis that the CM-Pf plays a pivotal role in goal-
oriented behavior selection. We here set out to test this hypothesis with electrophysiological
recordings from patients with electrodes implanted in CM-Pf for deep brain stimulation (DBS)
treatment of chronic neuropathic pain. Six patients participated in (1) an auditory three-class
oddball experiment, which required a button press to target tones, but not to standard and deviant
tones and in (2) a multi-speaker experiment with a target word that required attention selection
and a target image that required response selection. Subjects showed transient neural responses
(8-15 Hz) to the target tone and the target word. Two subjects additionally showed transient
neural responses (15-25 Hz) to the target image. All sensory target stimuli were related to an
internal goal and required a behavior selection (attention selection, response selection). In group
analyses, neural responses were greater to target tones than deviant and standard tones and to
target words than other task-relevant words that did not require attention selection. The transient

1 these authors contributed equally

2 these authors contributed equally

1
neural responses occurred after the target stimuli but prior to the overt behavioral response. Our
results demonstrate that in human subjects the CM-Pf is involved in signaling sensory inputs
related to goal-oriented selection of behavior.

Abbreviations
CM-Pf , centromedian-parafascicular complex; DBS, deep brain stimulation; FDR , false

discovery rate; LFP , local field potential

Keywords: intralaminar thalamus; attention; local field potentials; deep brain recordings; alpha

Introduction

It is vital for flexible goal-oriented behavior that we adaptively select appropriate behavior when
goal-relevant sensory input signals arrive. Such goal-oriented behavior may include overt or
covert attention selection and the selection of overt or covert goal-appropriate motor or non-
motor responses. Several studies have investigated goal-oriented attention and motor action
selection at the cortical level (Fan, 2014; Knight, 2007; Mesgarani and Chang, 2012) and the
cortical control of action selection in the basal ganglia (Jahfari et al., 2012). Evidence for an
involvement of the thalamic centromedian-parafascicular complex (CM-Pf) in processing of
relevant sensory events has been derived primarily from experimental studies in non-human
primates (Minamimoto and Kimura, 2002; Weigel and Krauss, 2004) and rodents (Kato et al.,
2011). However, little is known about the putative role of the CM-Pf in goal-oriented behavior
selection in humans.
Anatomically, the CM-Pf constitutes the major part of the intralaminar thalamus in
rodents and primates. It belongs to the higher-order thalamus and receives inputs from different
brainstem nuclei and cortical areas (Krout et al., 2002; Redgrave et al., 2010; Sadikot and
Rymar, 2009). The CM-Pf provides strong glutamatergic excitatory inputs to the striatum, its
main target structure (Doig et al., 2014; Lapper and Bolam, 1992; Nanda et al., 2009). The CM-
Pf additionally provides excitatory inputs to two other basal ganglia structures, the pallidum and

2
subthalamic nucleus, as well as to the cerebral cortex (Lanciego et al., 2009). The excitatory
input to the striatum seems to play an essential role in behavioral flexibility (Brown et al., 2010),
goal-directed motor action (Bradfield et al., 2013) and attention shifting (Ding et al., 2010). A
possible mechanism by which this thalamo-striatal input is modulating cortico-striatal-thalamo-
cortical processing is the synchronization of networks of cortical neurons. By synchronizing
multiple cortical regions, cortical information transmission may be controlled (Saalmann, 2014).
Earlier studies in non-human primates suggested a role of the CM-Pf in attentional
orienting (Minamimoto and Kimura, 2002) and processing of behaviorally relevant stimuli from
different sensory modalities, particularly audition (Matsumoto et al., 2001). A goal-oriented
component in CM-Pf signaling is suggested by findings in non-human primates showing that the
CM-Pf responds strongest to unexpected, undesired stimuli, which nevertheless require an action
linked to an internal goal (Minamimoto et al., 2014, 2005). In mice, the PF is a homologue of
the human CM-Pf (Groenewegen and Berendse, 1994), and disruption of the pathway from the
PF to the dorsolateral striatum leads to a decrease in correct button presses to visual stimuli in a
two-choice reaction time task (Kato et al., 2011). These results provide support for the notion
that the CM-Pf is important for goal-oriented response selection to behaviorally relevant stimuli.
The CM-Pf’s putative role in goal-oriented overt and covert selection of behavior is also in
accordance with its reciprocal connections with the motor and premotor cortices and cortical
areas of the attention and cognitive control networks (frontal eye field, prefrontal and anterior
cingulate cortices; (Sadikot and Rymar, 2009).
Very few studies have investigated the functional role of the CM-Pf in humans due to its
relative inaccessibility in the intralaminar thalamus. However, humans can perform much more
complex cognitive tasks and require a negligible amount of training compared to any animal
model. Interestingly, a PET study found greater regional blood flow in the intralaminar thalamus
when subjects were performing an attention-demanding task compared to a relaxed awake state
(Kinomura et al., 1996). Similarly, the intralaminar thalamus was activated during covert shifts
in visual attention in a recent fMRI study (Hulme et al., 2010). The temporal resolution of fMRI
and PET is, however, on the order of seconds, which means that the temporal dynamics of neural
responses in the human CM-Pf could not be determined in these imaging studies. The CM-Pf has
been introduced as a target for DBS in the treatment of chronic pain both with regard to its
functional connectivity and the previous experience with radiofrequency lesioning (Andy, 1980;

3
Hariz and Bergenheim, 1995; Krauss et al., 2002; Sims-Williams et al., 2016; Weigel and
Krauss, 2004). DBS in CM-Pf for treatment of chronic pain appears to modulate the affective
and motivational dimensions of pain perception (Weigel and Krauss, 2004). In stroke patients,
CM-Pf lesions result in deficits on the Wisconsin Card Sorting Test, which measures the ability
to flexibly adjust behavior to changing rules (Liebermann et al., 2013). Thus, there is some
evidence that the CM-Pf complex is relevant for attentional shifting and cognitive and behavioral
flexibility in humans.
Given these findings in non-human primates and limited evidence from human lesion and
imaging studies, we hypothesized that in humans the CM-Pf is involved in signaling the
occurrence of sensory events that are relevant for a current goal and request overt or covert
behavior selection, for example covert attention selection or response selection. In our study we
set out to characterize the signatures of CM-Pf responses with respect to this hypothesis.
Therefore, direct electrophysiological recordings were obtained in patients with CM-Pf DBS
electrodes while they performed tasks in which sensory cues signaled the need for behavior
selection to achieve the task goal. As tasks we selected a three-class auditory oddball and a
multi-speaker paradigm (Mesgarani and Chang, 2012; Polich, 2007). In oddball paradigms,
subjects are presented with frequent standard stimuli and infrequent stimuli that differ in some,
typically relatively simple, feature (Sandmann et al., 2009; Verleger et al., 2016, 2014). Here, a
three-class auditory oddball paradigm was used, in which subjects were presented with two
infrequent tones designated target and deviant tone, which differed in their frequency from a
frequent standard tone (Polich, 2007). The goal was to press a button only after target tones and
not after deviant and standard tones. Therefore, only the target tones required an overt response
selection. Typically, this task is used to measure attentional processes with EEG, which shows
enhanced responses around 300 ms after the infrequent stimuli. In general, this enhanced
deflection over fronto-parietal cortical regions reflects higher-level processing such as the
evaluation of a task-relevant event and updating of working memory (Kok, 2001; Verleger,
2008, 1997; Vogel et al., 1998). In the multi-speaker task two realistic speech streams were
presented simultaneously. The goal was to correctly report two task words from a target speaker
stream. The relevant speaker stream was indicated by a target word uttered by the target speaker.
The two task words appeared later in the target speaker stream and had to be reported after a
visual target indicated to the subjects that they should now press the buttons corresponding to the

4
two task words. This task required a covert attention selection after the target word, but not after
the two task words, and overt response selection after the target image. Moreover, using a visual
target to cue the response selection allowed us to investigate whether the CM-Pf responds to both
auditory and visual cues for goal-oriented behavior selection. Thus, the oddball and multi-
speaker task consist of target stimuli, which require different forms of behavior selection
(attention selection, response selection) and other sensory events that do not require a behavior
selection.
The reviewed studies suggest a behavior selection related plus a goal oriented component
in CM/Pf responses to external stimuli. Therefore, we expected that CM/Pf responds to all target
stimuli (target tone, target word, target image) because these were the sensory events that
required a behavior selection. We also expected neural responses to target stimuli to be greater
than responses to other task events which do not require a behavior selection. In particular, we
expected greater responses to the target tone compared to the deviant and standard tone and to
the target word compared to the two task words.

Materials and methods

Subject information
Informed consent was obtained from all subjects under the auspices of the ethics committee of
Hannover Medical School. Six patients with chronic neuropathic pain (2 women, age range 25-
57, mean age 48, SD 12 years; five subjects were right handed and one subject was left handed)
participated in an auditory oddball experiment and a multi-speaker experiment. The patients
were implanted with quadripolar DBS electrodes in the right CM-Pf (S1-S3, S6), the left CM-Pf
(S5) or bilateral CM-Pf (S4) for treatment of their pain syndromes. A summary of the clinical
background for the individual patients is provided in Table 1.

Electrode implantation and recording

Patients were implanted unilaterally, and in one case bilaterally, with quadripolar DBS electrodes
(Medtronic 3387, Minneapolis, MN, USA) both in the CM-Pf and the somatosensory thalamus
(ventral posterolateral or ventral posteromedial depending on pain distribution) guided by
stereotactic surgery as outlined in detail previously (Weigel and Krauss, 2004). DBS electrodes

5
were longitudinally spaced at distances of 1.5 mm (1.27 mm diameter, 1.5 mm length). A
stereotactic head frame was attached to the patients head under local anesthesia. Based on
stereotactic computer tomography (CT) imaging combined with a preoperative magnetic
resonance image (MR), the anterior (AC) and posterior (PC) commissures were identified to
calculate coordinates for the targets. Post-surgical stereotactic CT scans were performed to
document placement of DBS electrodes. Electrodes were externalized for several days for test
stimulation of both targets concerning its impact on pain. Thereafter, the target with the best
response was chosen for chronic stimulation or in case no response was achieved electrodes were
removed. The appropriate location of electrodes could be confirmed by the higher threshold for
electric stimulation induced sensory phenomena in the CM-Pf as compared to the somatosensory
thalamus.

Experimental design, task and stimuli

During both experiments, subjects were seated in a chair facing a video monitor (ViewSonic
VG930m, 1280x1024 resolution, 60 Hz frame rate) at a distance of approx. 64 cm. Sounds were
presented with two loudspeakers, one on each side of the video monitor at a volume that was
comfortable for the subjects and ensured comprehension.
On average, within three days after electrode placement, patients performed the auditory
three-class oddball experiment and the multi-speaker experiment. Four patients participated in
the two experiments in one session. Two patients were invited for a second session to participate
in the multi-speaker experiment. In the oddball paradigm (Polich, 2007), 700 auditory stimuli
were presented. Stimuli consisted of one frequent standard tone (900 Hz; 72%) and two
infrequent tones, either being a target tone (1200 Hz, 14%) or a deviant tone (600 Hz, 14%). The
experimental design of the oddball paradigm is summarized in Fig. 1A. Tones were presented for
62 ms, with a variable interstimulus interval between 800 and 1100 ms. The subjects were
instructed to press a button on the keyboard with their right index finger as soon as they heard
the target tone. Note that stimulus probabilities were matched between the deviant and target
tone, but only the target tone required a change in behavior. The oddball experiment took 12
minutes (2 blocks of 6 minutes each).
In the multi-speaker paradigm, an adapted German version of the Coordinate Response
Measure Speech Corpus (Bolia et al., 2000) was used. This paradigm was selected to investigate

6
CM-Pf responses during close-to-natural continuous speech stimulation with target words that
required behavior selection (attention selection) and other task relevant words (task words) that
did not require behavior selection. This paradigm allowed us to additionally introduce visual
targets that required a behavior selection (response selection). The trial structure of the multi-
speaker paradigm is depicted in Fig. 1B. All sentences started with the expression Los geht’s
(Let’s go) followed by a name (Goethe, Lessing, Schiller, Heine), one of four action expressions
with very similar meanings such as drücke gleich (press now) and a color number combination
such as grün und drei (green and three; Table 2). This structure resulted in sentences such as Los
geht’s Goethe entscheide dich für blau und zwei (Let’s go Goethe you choose blue and green).
Importantly, on each experimental trial two sentences were presented simultaneously and
binaurally, one by a female speaker and one by a male speaker. Before each experimental block,
which consisted of 24 sentence pairs, a name was presented in print on the video screen. The
name indicated the target speaker that should be attended in the multi-speaker stream because
this speaker would later utter the task-relevant color and number word. Therefore, the name
spoken by the target speaker on each trial is designated auditory target word in the sentence. The
irrelevant distracting speaker would always utter a different name (e.g. in the Goethe block, the
distracting speaker always uttered the name Lessing). After presentation of the sentences, an
image of a keyboard appeared on the screen. This target image indicated to the subject that she
could now press the buttons on the keyboard that corresponded to the color and the number word
that the target speaker, who had uttered the target word, had pronounced. Each sentence
contained one of four color words (green, blue, red, white) and one of four number words (two,
three, four, five) and these words differed between the two speakers on each trial. Subjects could
start the next trial by pressing the space bar. The target image stayed on the computer screen
until the subject pressed the space bar, then a fixation cross appeared on the screen and stayed on
the screen until the next target image appeared indicating the response interval. All sentences
were root-mean-square (RMS) normalized to the same volume level. Subjects were presented 96
trials in total. Note that only the target word and the target image required a behavior selection
(covert orienting of attention towards the target stream and overt manual response, respectively).
The multi-speaker experiment took approximately 20 minutes (4 blocks of 5 minutes each). As
the multi-speaker experiment was self-paced, the duration slightly differed between blocks and
patients.

7
Data acquisition
In both experiments, signals were sampled continuously at a rate of 512 Hz, amplified
10000-fold using a Digitimer D360 (Digitimer Ltd., Welwyn Garden City, Hertfordshire, UK)
and digitized through a 1401 A-D converter onto a computer using the Spike2 software (both
Cambridge Electronic Design, UK). Signals were recorded using a 50 Hz notch filter, a high-
pass filter with a cutoff frequency of 0.5 Hz and a low-pass filter with a cutoff frequency below
Nyquist frequency. Following previous studies (Herrojo Ruiz et al., 2014; Schroll et al., 2015;
Zaehle et al., 2013), LFPs were recorded or re-referenced bipolarly for further analysis. The
oddball experiment was directly recorded with bipolar referencing, while the multi-speaker
experiment was recorded with a reference electrode on the nose to allow flexible subsequent
bipolar re-referencing. Re-referencing was done after visual artifact rejection. Each contact was
referenced or re-referenced to its neighboring electrode contact (three bipolar channels per
hemisphere: 0-1, 1-2, 2-3) to maximize spatial selectivity and to reduce effects of volume
conduction from distant sources. Electrode impedances were below 5 kΩ during the whole
recording session.

Electrophysiological data analysis

The electrophysiological data were analyzed in MATLAB (MathWorks Inc. Natick, MA) using
the open source toolbox FieldTrip (Oostenveld et al., 2011) and customized scripts. Group
statistics were performed with SPSS (Version 23). The data were epoched with respect to
different stimulus onsets (target tone, standard tone, deviant tone, target word, task word I, task
word II, target image). Visual inspection of the data led to the rejection of 9% of trials in the
multi-speaker paradigm and no trials in the oddball paradigm. Additionally, all remaining trials
with button presses < 250 ms with respect to the target image were removed for all analyses on
the target image responses (13% of trials). This was done to ensure that only trials were included
where the subject waited for the visual cue to make the behavioral selection. To calculate the
mean latency of the button presses in the multi-speaker task, only the 1st button press of each trial
and only button presses with latencies >= 250 ms with respect to the target image were included.
For each subject one bipolar montage was selected for data analysis, which included
either contacts 0-1, 1-2, or 2-3. For all subjects, except S4, the identical montage was selected for

8
analyses of the oddball and multi-speaker data. S4 showed a clear response at a single right CM-
Pf contact in the oddball task (2-3) and a single left CM-Pf contact (0-1) in the multi-speaker
task. Montage selection criteria are described in more detail in the statistical analysis section.
For the analysis of the neural responses, the data were transformed to time-frequency
space using a single Hanning taper with a frequency dependent window length of four cycles
followed by Fourier transformation. The taper was applied in time steps of 10 ms and frequency
steps of 1 Hz to a frequency window from 3 to 40 Hz. Power values (squared analytic amplitude)
were calculated for each data point in the trial window. For the oddball task, individual baselines
were calculated for the different tone conditions (target, deviant, standard) at single trial level
before the onset of the auditory stimuli over the time window from -500 to 0 ms. We used the
same baseline from -500 to 0 ms (minimum onset across the two speaker streams) for the target
word, the target image and the two task words in the multi-speaker task. The average power over
the baseline window was used to calculate percent change for every data point relative to the
baseline at the individual trial level. Percent change values were then averaged across trials for
each time point. Our analyses focused on neural responses up to 40 Hz to speech and non-speech
stimuli because previous electrophysiological studies have shown strong responses in this
frequency range in subcortical structures (Brittain and Brown, 2014; Huang et al., 2016; Wróbel,
2000).

Statistical analysis
First, the time-frequency amplitude maps from 3 to 40 Hz masked at p < 0.05 (t-test to baseline,
uncorrected) were investigated at the three electrode montages for each subject. More
specifically, the average responses to the three tones (target, deviant, standard) from the oddball
task, the average responses to the target word and the two task words from the multi-speaker task
and the responses to the target image from the multi-speaker task were investigated. Based on the
greatest responses in any frequency in these three time-frequency amplitude maps, a single
montage was selected for each subject for further analysis. Only for S4 different montages were
used across the experiments, because this subject showed a response at a single right montage in
the oddball task and a single left montage in the multi-speaker task. Next, frequency bands for
further analysis were selected based on the most consistent responses across subjects. The
frequency range from 8-15 Hz was selected for the auditory targets (target tone, target word) and

9
the frequency range from 15-25 Hz for the target image. We also tested slight variations of the
selected frequency bands. This did not change the results we report. The time-frequency
responses for the three target stimuli are depicted in Fig. 2.
In the oddball experiment, to test how the neural responses were modulated by the tone
condition (target tone, deviant, standard), the maximum peak was identified based on the average
over all conditions separately for each subject. The mean amplitude over trials for each condition
and subject was calculated within a time window of ±50 ms around this maximum. Group
analyses using non-parametric one-factorial Friedman test and post-hoc Wilcoxon signed-rank
test were performed. To test for differences in pre-stimulus alpha, we averaged the baseline from
-500 to 0 ms for the alpha frequency range (8-15 Hz) and performed a repeated measures
ANOVA with the factor tone condition (target, deviant, standard) at the single subject level. The
Greenhouse-Geiser correction was applied to compensate for violations of the sphericity
assumption. There were no significant differences between pre-stimulus signals (range of F-
values: 0.425 – 2.253, range of p-values: .063 – .632).
As a control for a potential influence of response modality on the CM-Pf responses, one
subject (S5) participated in three versions of the oddball experiment, which only differed in the
task the subject had to perform. In the standard version, the subject pressed a button whenever a
target tone occurred, in the first control version, the subject silently counted all target tones, and
in the second control version, the subject pressed a button and silently counted all target tones.
The peak responses for a time window of ±50 ms around the average maximum within each
version of the task were extracted at the single trial level, and, given that baseline power values
across trials were normally distributed, a two-way independent ANOVA and post-hoc
independent t-tests were performed.
The same statistical group analyses as for the main oddball experiment were performed
for the multi-speaker experiment to determine how specific the transient neural responses were
to target words that require a goal-oriented behavior selection (attention selection). We compared
the responses to the target words with the responses to the first and second task words. The
neural responses to the target image were not compared in a group analysis because only S1 and
S2 showed clear transient responses after the target image (Fig. 2).
At the single subject level, we wanted to determine whether the different targets (target
tone, target word, target image) elicited significant amplitude increases compared to baseline in

10
the selected frequency ranges and when responses to the targets peaked. We performed running
t-tests on the data segments from 0-600 ms for the individual subjects for both experiments in
order to compare the responses to the target tone, the target word and the target image with
respect to baseline. The running t-tests were false discovery rate (FDR) corrected at q = 0.01
(Benjamini and Hochberg, 1995) using the method of Groppe and colleagues (Groppe et al.,
2011).

Results

Behavioral performance
In the auditory oddball paradigm, S1, S2, S5 and S6 showed a high accuracy of at least 95%
correct responses. S3 and S4 had slightly lower hit rates of 71% and 81% correct trials,
respectively. The false alarm rate to standard or deviant tones was below 10% for all subjects. In
the multi-speaker paradigm, subject S1 and S2 showed high accuracy with correct number and
color button presses on 95% of trials. S6 showed an accuracy of 93%, while S3-S5 showed
somewhat lower performance with 83%, 83% and 89% correct responses, respectively. These
results indicate that the subjects were able to perform both tasks well.

Time-frequency responses in the CM-Pf to the target tone, target word and target image
The time-frequency responses were analyzed for each individual subject at the selected CM-Pf
contacts. Clear transient responses were detected to the target tone in the oddball task (Fig. 2A)
and the target word in the multi-speaker task (Fig. 2B) for all six subjects in the 8-15 Hz
frequency range. The magnitudes of the neural responses to the target word were less strong
compared to the responses to the target tone. Only two subjects, S1 and S2, showed strong
transient responses to the target image in the 15-25 Hz range (Fig. 2C). The amplitudes of the
neural responses reported in the following analyses were calculated as the average amplitude
over these frequency bands.

Neural responses to the target tone and target word compared to other task events
Next, we performed group analyses for those target stimuli that elicited transient neural
responses in all six subjects (target tone, target word). We aimed to determine whether the CM-
Pf shows strong transient neural responses specifically to sensory events that require overt or

11
covert goal-oriented behavior selection. Figure 3A shows the time courses of the neural
responses (8-15 Hz) to the target tone (purple curves), deviant tone and standard tone. Figure 3B
shows the neural responses (8-15 Hz) to the target word (red curves), task word I and task word
II. The target stimuli, which required a behavior selection in contrast to the other task events,
elicited transient responses in the CM-Pf.
We first tested whether the transient neural responses to the target tone in the oddball task
were significantly greater than to the standard tones and, particularly, to the deviant tones. In the
oddball experiment, the default response mode is no-response. In the case of a target tone, the
subject needs to adapt this behavior and overtly respond with a button press. Therefore, we
expected a greater neural response in the CM-Pf to target tones compared to other tones, as they
require a goal-oriented behavior selection. In a group analysis using a non-parametric two-
factorial Friedman test, we first determined that there were differences in response peaks across
the three conditions (χ2(2) = 12.00, p = .002). Post-hoc Wilcoxon signed-rank tests between the
different conditions revealed that the response peaks to target tones were greater compared to
both deviant tones and standard tones (Z = 2.20, p = .028; Fig. 3A). Furthermore, deviant tones
elicited a stronger response compared to standard tones (Z = 2.20, p = .028). As the target tone is
the only tone that required a button press, we wanted to determine in a control experiment with
one subject whether the strong response observed to target tones is only present when an overt
motor response is required or whether it also occurs when a covert behavioral response is
required. Therefore, S5 performed the oddball paradigm with three versions of the response:
button press, silent counting and button press + silent counting. When we performed a 3 x 3
ANOVA with factors condition (target tone, deviant, standard) and task (button press, silent
counting, button press + silent counting) on the single trial data from S5, we observed a main
effect of condition (F(2,387) = 12.626, p < .001), but no main effect of task (F(2,387) = 1.103, p
= .333) and no interaction effect (F(4,387) = 1.163, p = .327). For post-hoc tests, the data were
collapsed across tasks and it was again found that neural responses to target tones were greater
compared to both deviant tones (t(43) = 3.925, p < .001) and standard tones (t(43) = 5.231, p <
.001). The response to deviant tones was also larger compared with the response to standard
tones (t(43) = 2.5, p = .016). Together, these results indicate that simple frequency modulated
tones that signal the requirement for a goal-oriented behavior selection (button press, or
counting, or both) elicit the strongest responses in CM-Pf irrespective of the type of behavior.

12
The comparison between target tones and deviant tones indicates that tone probability modulated
CM-Pf responses much less than the behavioral significance of the tone. As the oddball
experiment with different tasks was only performed by a single subject, these results should be
interpreted with caution.
Next, we wanted to determine whether the CM-Pf response is specific to simple auditory
stimuli or if it is also elicited in tasks that use more complex stimuli and require goal-oriented
covert behavior selection, more specifically attention selection. We compared peak responses in
the multi-speaker experiment between the target word and the two task words (color and number
word). We expected stronger transient neural responses in the CM-Pf to target words compared
to task words because the target words signaled to the subjects that they should focus their
attention on the respective speaker stream. In a group analysis across all six subjects using a non-
parametric one-factorial Friedman test, we found significant differences across the three
conditions target word, task word I and task word II (χ2(2) = 9.00, p = .011). Post-hoc Wilcoxon
signed-rank tests revealed greater responses to the target word than to task word I and task word
II (Z = 2.20, p = .028; Fig. 3B). No significant differences were observed between the responses
to the two task words (Z = 0.31, p = .753).

Timing of neural responses to the different targets at single subject level

Based on the frequency response profiles to the three targets (Fig. 2), time courses of the neural
responses were next compared to baseline at the single subject level. Peak latencies of the neural
responses and reaction times are provided in Table 3. All six subjects showed significant
increases in CM-Pf activity to the target tone compared to baseline (running t-tests, 0-600 ms,
FDR corrected at q = 0.01), with peak latencies between 164 and 193 ms. Four subjects showed
significant increases in CM-Pf activity to the target word compared to baseline with peak
latencies between 130 and 300 ms. Subjects S1 and S2 showed significant increases in CM-Pf
activity to the target image compared to baseline, with peak latencies of 340 and 550 ms. The
average neural responses always peaked after target presentation and before the average reaction
times.

Discussion

13
Previous experimental research suggests a role of the CM-Pf in attention and response selection.
Our study demonstrates that in humans, the CM-Pf responds strongly to sensory events related to
a current goal that require an overt or covert behavior selection in different target scenarios.
Specifically, we found the strongest transient neural responses to simple auditory target
tones that required a button press, auditory target words that initiated attention selection of a
speaker stream, and, in two subjects, visual target images that initiated response selection. A
study in non-human primates similarly showed increased neural responses in the CM-Pf after a
visual target that required spatial attentional orienting as well as after a visual target that cued a
button release (Minamimoto and Kimura, 2002). In this study, inactivation of the CM-Pf led to
increased reaction times specifically to validly cued targets in the contralateral visual field. This
further supports the hypothesis that the CM-Pf plays a key role in goal-oriented behavior
selection as also seen during the oddball paradigm with enhanced neural responses to salient
target tones compared with frequent standard tones. Indeed, the oddball paradigm is frequently
used in EEG studies to investigate attention-related modulation of cortical activation. There, the
paradigm is known to reliably elicit enhanced responses to salient target events that are
considered to reflect the allocation of attention that is required to evaluate a task-relevant event
and to update working memory accordingly (Bledowski et al., 2004; Donchin and Coles, 1988;
Kok, 2001; Verleger, 2008, 1997).
In a study in non-human primates, it was shown that the CM-Pf does not simply respond
strongest to a stimulus that signals a large reward, but that the CM-Pf seems to participate in
shifting the behavior from the bias towards large rewards to requested small-reward actions
(Minamimoto et al., 2005). In the oddball experiment, there was a clear bias towards no
response, thus the response increase to target tones in our study is in line with the idea that the
CM-Pf “signals the discrepancy between internal pre-action bias and external demand”
(Minamimoto et al., 2014). However, the response profile of the CM-Pf is more complex. In the
multi-speaker experiment, there was no inherent bias in the paradigm towards one of the
speakers (female, male) and in the oddball paradigm, deviant tones, which did not require a shift
from the pre-action bias, elicited larger responses than standard tones. Our results indicate that
the CM-Pf responds strongest whenever the need for a goal-oriented behavior selection is
signaled by an external stimulus. This can include various overt and covert behaviors as for
example attention shifting, button presses, and covert counting of targets.

14
 Task-relevant (task words in the multi-speaker paradigm) and irrelevant (standard and
deviant tones in the oddball paradigm) stimuli that did not require a direct behavior selection
induced significantly weaker CM-Pf responses than the target stimuli in our experiments. This
finding is important, because it shows that task relevance alone does not induce a strong response
in the CM-Pf, but that an overt or covert behavior selection within a task context is required to
initiate strong responses in the CM-Pf. A selection or shift in behavior was also necessary in a
study by Liebermann and colleagues (2013). They examined the performance of eight stroke
patients with lesions in the CM-Pf on the Wisconsin Card Sorting Test. In this test, new rules
have to be identified, maintained for some time and behavior needs to be adapted accordingly.
The patients with lesions in the CM-Pf showed deficits on several dimensions of the Wisconsin
Card Sorting Test. The authors proposed that the CM-Pf may contribute to “shifting of cognitive
sets by virtue of its dense connections with the striatum” (Liebermann et al., 2013).
Finally, we found that in cases when the target stimulus was followed by a button press
(target tone, target image), the transient neural responses occurred before the button press. This
is in line with findings in animals that suggest CM-Pf involvement in response selection, but not
response execution (Kato et al., 2011; Minamimoto et al., 2005). In the single subject that
participated in an oddball experiment with three different tasks (button press, silent counting and
button press + silent counting), the neural response increase to the target tones did not
significantly differ between the tasks. This suggests that the CM-Pf is always activated when a
sensory stimulus requires a goal-oriented overt or covert behavior selection, and not just when a
response selection is required that results in an overt response such as a button press. This
interpretation is further supported by the transient neural response increase we found after the
target words, which initiated a covert attention selection. Such attention selection in a similar
paradigm has previously been shown in the human auditory cortex (Mesgarani and Chang,
2012). A major difference between the neural responses in the auditory cortex and the CM-Pf is,
however, that the CM-Pf does not seem to code the modality of the presented sensory stimulation
and that it does not show a response that is sustained over the whole interval of attention
selection. Instead, it shows a transient response when a stimulus occurs that signals a required
behavior selection.
The goal-oriented behavior selection signal in the CM-Pf seems to be modality-
independent because we observed it to simple auditory tones, auditory words and visual images,

15
albeit only in two subjects to visual images. Indeed, previous studies in behaving non-human
primates noted strong neural responses to task-relevant and unexpected visual, auditory and
somatosensory stimuli, suggesting strong CM-Pf activation by unpredicted, salient stimuli
(Matsumoto et al., 2001; Minamimoto et al., 2005; Minamimoto and Kimura, 2002). The CM-Pf
responses to the different target stimuli slightly differed in the frequency range of the responses,
which may be related to the difference in modality (auditory, visual) and complexity (simple
tone, word, image) between the target stimuli as well as the nature of the different behavioral
tasks (selecting a task-relevant speaker, pressing a button only after target tones, adaptively
selecting correct buttons). However, all targets elicited CM-Pf responses in the alpha to beta
range. Other thalamic nuclei (pulvinar, LGN) have been shown to synchronize with cortical
brain regions in the alpha to beta range and this synchronization is modulated by attention
(Bastos et al., 2014; Saalmann et al., 2012).
The CM-Pf selection signal may impact cortical and basal ganglia processing through its
efferent connections with the striatum as well as the prefrontal and (pre)motor cortex (Parent and
Parent, 2005; Sadikot et al., 1992; Smith and Parent, 1986). Neural responses in the CM-Pf may
be strongly influenced by activity in subcortical and cortical structures through its many afferent
connections. The CM-Pf receives input from motor, premotor and somatosensory cortex, from
prefrontal cortex, including the frontal eye field and anterior cingulate cortex, and from several
subcortical regions including the basal ganglia, the reticular formation and the superior colliculus
(Krout et al., 2002; May et al., 2007; Van der Werf et al., 2002). The superior colliculus is one
major input structure to the CM-Pf which receives both auditory and visual input (Gaese and
Johnen, 2000; Harting et al., 1992). It has been shown that inhibition of the superior colliculus
impairs covert and overt target selection (Lovejoy and Krauzlis, 2010; Song et al., 2011). Thus,
the superior colliculus may play an important role in the selection of simple stimuli, such as
target tone and target image. The CM-Pf, in turn, strongly projects to the basal ganglia (Sadikot
et al., 1992; Smith and Parent, 1986), which play a major role in action and motor program
selection (McHaffie et al., 2005) and, therefore, flexible behavior. Thus, the colliculo-thalamo-
basal ganglia-collicular loops may be a crucial circuitry in overt and covert behavior selection
(Fisher and Reynolds, 2014; Krauzlis et al., 2014; McHaffie et al., 2005; Redgrave et al., 1999).
In this circuitry, the superior colliculus may transmit a signal on salient stimuli to the CM-Pf. At
the same time, endogenous behavioral relevance, which was prelearned in our experiments, may

16
be provided to the CM-Pf via cortical afferents and also indirectly via cortical projections to the
superior colliculus (Fries, 1984). Based on anatomical and functional studies in animals, the
goal-oriented overt and covert behavior selection signal in the CM-Pf impacts on basal ganglia
processing via the CM-Pf’s strong efferent connectivity with the striatum. Influences of
excitatory inputs from the CM-Pf to the striatum on behavioral flexibility and learning of new
action-outcome contingencies have previously been shown in rats (Bradfield et al., 2013; Brown
et al., 2010). The observed behavioral effects in the two latter studies can also be explained by an
impairment in goal-oriented behavior selection and, therefore, provide a likely underlying
mechanism on how CM-Pf affects basal ganglia processing in our tasks. Goal-oriented behaviors
may then be selected and executed via cortico-striato-thalamo-cortical circuitry.
The CM-Pf may also play a role in the cognitive deficits observed in patients with
Parkinson’s disease. Apart from the motor deficits observed in patients with Parkinson’s disease,
many patients show cognitive deficits in set-shifting and dual-tasking skills (Lowit et al., 2005).
Regarding the putative function of CM-Pf (signaling sensory cues for goal-oriented behavior
selection) and basal ganglia (filter for information and action selection) with their close
connectivity, the question arises to which degree cognitive deficits in Parkinson’s disease might
not only be associated to a decrease in dopamine and thus basal ganglia dysfunction, but also to
malfunction of the CM-Pf. Postmortem studies of Parkinson’s disease patients have shown
severe cell loss in the CM-Pf (Henderson et al., 2000a, 2000b). Furthermore, behavioral studies
with Parkinson’s patients have shown a deficit in goal-oriented cognitive control, particularly in
instances of “internal cueing”, that is when a goal had to be kept in memory (Brown and
Marsden, 1988; Stolwyk et al., 2005). In the study by Brown and Marsden (1988) for example,
reaction times in Parkinson’s disease patients on a Stroop task most strongly increased when the
feature to which the subjects had to respond (ink, word) was presented not before each trial but
before the entire experimental block. Similarly, in our multi-speaker paradigm, subjects had to
remember the target word, which identified the task-relevant speaker over an entire experimental
block. In the auditory oddball paradigm subjects had to remember the target tone over the entire
experiment. The CM-Pf may play a key role in such goal-oriented “internal cueing” situations,
and CM-Pf malfunction may underlie deficits related to goal-oriented cueing in Parkinson’s
disease.

17
 In summary, based on our and previous findings, we propose that the CM-Pf provides
neural signals that can trigger overt or covert behavior selection. The CM-Pf signal is initiated by
changing sensory input and depends on an internal goal (e.g. selecting a task-relevant speaker,
silently counting, pressing a button only after target tones, adaptively selecting correct buttons).
The CM-Pf may contribute to goal-oriented selection of behavior via both colliculo-thalamo-
basal ganglia-collicular loops and cortico-striato-thalamo-cortical loops. Future studies are
needed to investigate the functional interaction between the different subcortical and cortical
regions in goal-oriented overt and covert behavior selection and behavior execution.

Acknowledgments
We thank all patients and their families, whose cooperation made this research possible. We also
thank Marc René Schädler and Bernd T. Meyer for phonetic characterization of the sentences for
the multi-speaker experiment and Avgusta Shestyuk for useful comments. This work was
supported by the Deutsche Forschungsgemeinschaft (Cluster of Excellence 1077 “Hearing4all”).

18
References

Andy, O.J., 1980. Parafascicular-center median nuclei stimulation for intractable pain and
dyskinesia (painful-dyskinesia). Appl. Neurophysiol. 43, 133–144.
Bastos, A.M., Briggs, F., Alitto, H.J., Mangun, G.R., Usrey, W.M., 2014. Simultaneous
recordings from the primary visual cortex and lateral geniculate nucleus reveal rhythmic
interactions and a cortical source for γ-band oscillations. J. Neurosci. Off. J. Soc.
Neurosci. 34, 7639–7644. doi:10.1523/JNEUROSCI.4216-13.2014
Benjamini, Y., Hochberg, Y., 1995. Controlling the False Discovery Rate: A Practical and
Powerful Approach to Multiple Testing. J. R. Stat. Soc. Ser. B Methodol. 57, 289–300.
Bledowski, C., Prvulovic, D., Goebel, R., Zanella, F.E., Linden, D.E.J., 2004. Attentional
systems in target and distractor processing: a combined ERP and fMRI study.
NeuroImage 22, 530–540. doi:10.1016/j.neuroimage.2003.12.034
Bolia, R.S., Nelson, W.T., Ericson, M.A., Simpson, B.D., 2000. A speech corpus for multitalker
communications research. J. Acoust. Soc. Am. 107, 1065–1066.
Bradfield, L.A., Bertran-Gonzalez, J., Chieng, B., Balleine, B.W., 2013. The thalamostriatal
pathway and cholinergic control of goal-directed action: interlacing new with existing
learning in the striatum. Neuron 79, 153–166. doi:10.1016/j.neuron.2013.04.039
Brittain, J.-S., Brown, P., 2014. Oscillations and the basal ganglia: motor control and beyond.
NeuroImage 85 Pt 2, 637–647. doi:10.1016/j.neuroimage.2013.05.084
Brown, H.D., Baker, P.M., Ragozzino, M.E., 2010. The parafascicular thalamic nucleus
concomitantly influences behavioral flexibility and dorsomedial striatal acetylcholine
output in rats. J. Neurosci. Off. J. Soc. Neurosci. 30, 14390–14398.
doi:10.1523/JNEUROSCI.2167-10.2010
Brown, R.G., Marsden, C.D., 1988. Internal versus external cues and the control of attention in
Parkinson’s disease. Brain J. Neurol. 111 ( Pt 2), 323–345.
Ding, J.B., Guzman, J.N., Peterson, J.D., Goldberg, J.A., Surmeier, D.J., 2010. Thalamic gating
of corticostriatal signaling by cholinergic interneurons. Neuron 67, 294–307.
doi:10.1016/j.neuron.2010.06.017
Doig, N.M., Magill, P.J., Apicella, P., Bolam, J.P., Sharott, A., 2014. Cortical and thalamic
excitation mediate the multiphasic responses of striatal cholinergic interneurons to
motivationally salient stimuli. J. Neurosci. Off. J. Soc. Neurosci. 34, 3101–3117.
doi:10.1523/JNEUROSCI.4627-13.2014
Donchin, E., Coles, M.G.H., 1988. Is the P300 component a manifestation of context updating?
Behav. Brain Sci. 11, 357–374. doi:10.1017/S0140525X00058027
Fan, J., 2014. An information theory account of cognitive control. Front. Hum. Neurosci. 8, 680.
doi:10.3389/fnhum.2014.00680
Fisher, S.D., Reynolds, J.N.J., 2014. The intralaminar thalamus-an expressway linking visual
stimuli to circuits determining agency and action selection. Front. Behav. Neurosci. 8,
115. doi:10.3389/fnbeh.2014.00115
Fries, W., 1984. Cortical projections to the superior colliculus in the macaque monkey: a
retrograde study using horseradish peroxidase. J. Comp. Neurol. 230, 55–76.
doi:10.1002/cne.902300106
Gaese, B.H., Johnen, A., 2000. Coding for auditory space in the superior colliculus of the rat.
Eur. J. Neurosci. 12, 1739–1752.

19
Groenewegen, H.J., Berendse, H.W., 1994. The specificity of the “nonspecific” midline and
intralaminar thalamic nuclei. Trends Neurosci. 17, 52–57.
Groppe, D.M., Urbach, T.P., Kutas, M., 2011. Mass univariate analysis of event-related brain
potentials/fields I: a critical tutorial review. Psychophysiology 48, 1711–1725.
doi:10.1111/j.1469-8986.2011.01273.x
Hariz, M.I., Bergenheim, A.T., 1995. Thalamic stereotaxis for chronic pain: ablative lesion or
stimulation? Stereotact. Funct. Neurosurg. 64, 47–55.
Harting, J.K., Updyke, B.V., Van Lieshout, D.P., 1992. Corticotectal projections in the cat:
anterograde transport studies of twenty-five cortical areas. J. Comp. Neurol. 324, 379–
414. doi:10.1002/cne.903240308
Henderson, J.M., Carpenter, K., Cartwright, H., Halliday, G.M., 2000a. Loss of thalamic
intralaminar nuclei in progressive supranuclear palsy and Parkinson’s disease: clinical
and therapeutic implications. Brain J. Neurol. 123 ( Pt 7), 1410–1421.
Henderson, J.M., Carpenter, K., Cartwright, H., Halliday, G.M., 2000b. Degeneration of the
centré median-parafascicular complex in Parkinson’s disease. Ann. Neurol. 47, 345–352.
Herrojo Ruiz, M., Rusconi, M., Brücke, C., Haynes, J.-D., Schönecker, T., Kühn, A.A., 2014.
Encoding of sequence boundaries in the subthalamic nucleus of patients with Parkinson’s
disease. Brain J. Neurol. 137, 2715–2730. doi:10.1093/brain/awu191
Huang, Y., Geng, X., Li, L., Stein, J.F., Aziz, T.Z., Green, A.L., Wang, S., 2016. Measuring
complex behaviors of local oscillatory networks in deep brain local field potentials. J.
Neurosci. Methods 264, 25–32. doi:10.1016/j.jneumeth.2016.02.018
Hulme, O.J., Whiteley, L., Shipp, S., 2010. Spatially distributed encoding of covert attentional
shifts in human thalamus. J. Neurophysiol. 104, 3644–3656. doi:10.1152/jn.00303.2010
Jahfari, S., Verbruggen, F., Frank, M.J., Waldorp, L.J., Colzato, L., Ridderinkhof, K.R.,
Forstmann, B.U., 2012. How preparation changes the need for top-down control of the
basal ganglia when inhibiting premature actions. J. Neurosci. Off. J. Soc. Neurosci. 32,
10870–10878. doi:10.1523/JNEUROSCI.0902-12.2012
Kato, S., Kuramochi, M., Kobayashi, K., Fukabori, R., Okada, K., Uchigashima, M., Watanabe,
M., Tsutsui, Y., Kobayashi, K., 2011. Selective neural pathway targeting reveals key
roles of thalamostriatal projection in the control of visual discrimination. J. Neurosci. Off.
J. Soc. Neurosci. 31, 17169–17179. doi:10.1523/JNEUROSCI.4005-11.2011
Kinomura, S., Larsson, J., Gulyás, B., Roland, P.E., 1996. Activation by attention of the human
reticular formation and thalamic intralaminar nuclei. Science 271, 512–515.
Knight, R.T., 2007. Neural Networks Debunk Phrenology. Science 316, 1578–1579.
doi:10.1126/science.1144677
Kok, A., 2001. On the utility of P3 amplitude as a measure of processing capacity.
Psychophysiology 38, 557–577.
Krauss, J.K., Pohle, T., Weigel, R., Burgunder, J.-M., 2002. Deep brain stimulation of the centre
median-parafascicular complex in patients with movement disorders. J. Neurol.
Neurosurg. Psychiatry 72, 546–548.
Krauzlis, R.J., Bollimunta, A., Arcizet, F., Wang, L., 2014. Attention as an effect not a cause.
Trends Cogn. Sci. 18, 457–464. doi:10.1016/j.tics.2014.05.008
Krout, K.E., Belzer, R.E., Loewy, A.D., 2002. Brainstem projections to midline and intralaminar
thalamic nuclei of the rat. J. Comp. Neurol. 448, 53–101. doi:10.1002/cne.10236
Lanciego, J.L., López, I.P., Rico, A.J., Aymerich, M.S., Pérez-Manso, M., Conte, L., Combarro,
C., Roda, E., Molina, C., Gonzalo, N., Castle, M., Tuñón, T., Erro, E., Barroso-Chinea,

20
 P., 2009. The search for a role of the caudal intralaminar nuclei in the pathophysiology of
Parkinson’s disease. Brain Res. Bull. 78, 55–59. doi:10.1016/j.brainresbull.2008.08.008
Lapper, S.R., Bolam, J.P., 1992. Input from the frontal cortex and the parafascicular nucleus to
cholinergic interneurons in the dorsal striatum of the rat. Neuroscience 51, 533–545.
Liebermann, D., Ploner, C.J., Kraft, A., Kopp, U.A., Ostendorf, F., 2013. A dysexecutive
syndrome of the medial thalamus. Cortex J. Devoted Study Nerv. Syst. Behav. 49, 40–49.
doi:10.1016/j.cortex.2011.11.005
Lovejoy, L.P., Krauzlis, R.J., 2010. Inactivation of primate superior colliculus impairs covert
selection of signals for perceptual judgments. Nat. Neurosci. 13, 261–266.
doi:10.1038/nn.2470
Lowit, A., Howell, P., Brendel, B., 2005. Cognitive Impairment in Parkinson’s Disease: Is It a
Unified Phenomenon? Brain Impair. Multidiscip. J. Aust. Soc. Study Brain Impair. 6,
191–204.
Matsumoto, N., Minamimoto, T., Graybiel, A.M., Kimura, M., 2001. Neurons in the thalamic
CM-Pf complex supply striatal neurons with information about behaviorally significant
sensory events. J. Neurophysiol. 85, 960–976.
May, A., Hajak, G., Ganssbauer, S., Steffens, T., Langguth, B., Kleinjung, T., Eichhammer, P.,
2007. Structural brain alterations following 5 days of intervention: dynamic aspects of
neuroplasticity. Cereb Cortex 17, 205–10.
McHaffie, J.G., Stanford, T.R., Stein, B.E., Coizet, V., Redgrave, P., 2005. Subcortical loops
through the basal ganglia. Trends Neurosci. 28, 401–407. doi:10.1016/j.tins.2005.06.006
Mesgarani, N., Chang, E.F., 2012. Selective cortical representation of attended speaker in multi-
talker speech perception. Nature 485, 233–236. doi:10.1038/nature11020
Minamimoto, T., Hori, Y., Kimura, M., 2005. Complementary process to response bias in the
centromedian nucleus of the thalamus. Science 308, 1798–1801.
doi:10.1126/science.1109154
Minamimoto, T., Hori, Y., Yamanaka, K., Kimura, M., 2014. Neural signal for counteracting
pre-action bias in the centromedian thalamic nucleus. Front. Syst. Neurosci. 8, 3.
doi:10.3389/fnsys.2014.00003
Minamimoto, T., Kimura, M., 2002. Participation of the thalamic CM-Pf complex in attentional
orienting. J. Neurophysiol. 87, 3090–3101.
Nanda, B., Galvan, A., Smith, Y., Wichmann, T., 2009. Effects of stimulation of the
centromedian nucleus of the thalamus on the activity of striatal cells in awake rhesus
monkeys. Eur. J. Neurosci. 29, 588–598. doi:10.1111/j.1460-9568.2008.06598.x
Oostenveld, R., Fries, P., Maris, E., Schoffelen, J.M., 2011. FieldTrip: Open source software for
advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell
Neurosci 2011, 156869.
Parent, M., Parent, A., 2005. Single-axon tracing and three-dimensional reconstruction of centre
median-parafascicular thalamic neurons in primates. J. Comp. Neurol. 481, 127–144.
doi:10.1002/cne.20348
Polich, J., 2007. Updating P300: an integrative theory of P3a and P3b. Clin Neurophysiol 118,
2128–48.
Redgrave, P., Coizet, V., Comoli, E., McHaffie, J.G., Leriche, M., Vautrelle, N., Hayes, L.M.,
Overton, P., 2010. Interactions between the Midbrain Superior Colliculus and the Basal
Ganglia. Front. Neuroanat. 4. doi:10.3389/fnana.2010.00132

21
Redgrave, P., Prescott, T.J., Gurney, K., 1999. The basal ganglia: a vertebrate solution to the
selection problem? Neuroscience 89, 1009–1023.
Saalmann, Y.B., 2014. Intralaminar and medial thalamic influence on cortical synchrony,
information transmission and cognition. Front. Syst. Neurosci. 8, 83.
doi:10.3389/fnsys.2014.00083
Saalmann, Y.B., Pinsk, M.A., Wang, L., Li, X., Kastner, S., 2012. The pulvinar regulates
information transmission between cortical areas based on attention demands. Science
337, 753–756. doi:10.1126/science.1223082
Sadikot, A.F., Parent, A., Smith, Y., Bolam, J.P., 1992. Efferent connections of the centromedian
and parafascicular thalamic nuclei in the squirrel monkey: a light and electron
microscopic study of the thalamostriatal projection in relation to striatal heterogeneity. J.
Comp. Neurol. 320, 228–242. doi:10.1002/cne.903200207
Sadikot, A.F., Rymar, V.V., 2009. The primate centromedian-parafascicular complex:
anatomical organization with a note on neuromodulation. Brain Res. Bull. 78, 122–130.
doi:10.1016/j.brainresbull.2008.09.016
Sandmann, P., Eichele, T., Buechler, M., Debener, S., Jäncke, L., Dillier, N., Hugdahl, K.,
Meyer, M., 2009. Evaluation of evoked potentials to dyadic tones after cochlear
implantation. Brain J. Neurol. 132, 1967–1979. doi:10.1093/brain/awp034
Schroll, H., Horn, A., Gröschel, C., Brücke, C., Lütjens, G., Schneider, G.-H., Krauss, J.K.,
Kühn, A.A., Hamker, F.H., 2015. Differential contributions of the globus pallidus and
ventral thalamus to stimulus-response learning in humans. NeuroImage 122, 233–245.
doi:10.1016/j.neuroimage.2015.07.061
Sims-Williams, H.P., Javed, S., Pickering, A.E., Patel, N.K., 2016. Characterising the Analgesic
Effect of Different Targets for Deep Brain Stimulation in Trigeminal Anaesthesia
Dolorosa. Stereotact. Funct. Neurosurg. 94, 174–181. doi:10.1159/000446608
Smith, Y., Parent, A., 1986. Differential connections of caudate nucleus and putamen in the
squirrel monkey (Saimiri sciureus). Neuroscience 18, 347–371.
Song, J.-H., Rafal, R.D., McPeek, R.M., 2011. Deficits in reach target selection during
inactivation of the midbrain superior colliculus. Proc. Natl. Acad. Sci. U. S. A. 108,
E1433-1440. doi:10.1073/pnas.1109656108
Stolwyk, R.J., Triggs, T.J., Charlton, J.L., Iansek, R., Bradshaw, J.L., 2005. Impact of internal
versus external cueing on driving performance in people with Parkinson’s disease. Mov.
Disord. Off. J. Mov. Disord. Soc. 20, 846–857. doi:10.1002/mds.20420
Van der Werf, Y.D., Witter, M.P., Groenewegen, H.J., 2002. The intralaminar and midline nuclei
of the thalamus. Anatomical and functional evidence for participation in processes of
arousal and awareness. Brain Res. Brain Res. Rev. 39, 107–140.
Verleger, R., 2008. P3b: towards some decision about memory. Clin. Neurophysiol. Off. J. Int.
Fed. Clin. Neurophysiol. 119, 968–970. doi:10.1016/j.clinph.2007.11.175
Verleger, R., 1997. On the utility of P3 latency as an index of mental chronometry.
Psychophysiology 34, 131–156.
Verleger, R., Grauhan, N., Śmigasiewicz, K., 2016. Is P3 a strategic or a tactical component?
Relationships of P3 sub-components to response times in oddball tasks with go, no-go
and choice responses. NeuroImage 143, 223–234. doi:10.1016/j.neuroimage.2016.08.049
Verleger, R., Metzner, M.F., Ouyang, G., Śmigasiewicz, K., Zhou, C., 2014. Testing the
stimulus-to-response bridging function of the oddball-P3 by delayed response signals and

22
 residue iteration decomposition (RIDE). NeuroImage 100, 271–280.
doi:10.1016/j.neuroimage.2014.06.036
Vogel, E.K., Luck, S.J., Shapiro, K.L., 1998. Electrophysiological evidence for a postperceptual
locus of suppression during the attentional blink. J. Exp. Psychol. Hum. Percept. Perform.
24, 1656–1674.
Weigel, R., Krauss, J.K., 2004. Center median-parafascicular complex and pain control. Review
from a neurosurgical perspective. Stereotact. Funct. Neurosurg. 82, 115–126.
doi:10.1159/000079843
Wróbel, A., 2000. Beta activity: a carrier for visual attention. Acta Neurobiol. Exp. (Warsz.) 60,
247–260.
Zaehle, T., Bauch, E.M., Hinrichs, H., Schmitt, F.C., Voges, J., Heinze, H.-J., Bunzeck, N.,
2013. Nucleus accumbens activity dissociates different forms of salience: evidence from
human intracranial recordings. J. Neurosci. Off. J. Soc. Neurosci. 33, 8764–8771.
doi:10.1523/JNEUROSCI.5276-12.2013

Figure Legends
Figure 1. Experimental setup. (A) Experimental design of the oddball paradigm. Target tones
(purple boxes) and deviant tones were presented with the same probability but only the target
tones required a response. (B) Trial structure of the multi-speaker paradigm. On each trial two
speaker streams (male and female) were presented simultaneously and binaurally to the subject.
Before each experimental block, a name was presented in print on the video screen, indicating
that the speaker who utters this name will also provide task-relevant information. Therefore, this
spoken name was designated target word (red box). The target word required a covert attention
shift to the target stream. The color and number words (task word I and II) were uttered later in
this speech stream. An image of a keyboard, designated target image (blue box) was presented at
the end of the sentence, indicating to subjects that they should now make the responses. Upon
appearance of the target image, the subject’s task was to press the buttons referring to the color
and number that was uttered in the target speech stream.

Figure 2. Time-frequency response amplitudes in CM-Pf to the different target stimuli. (A)
Responses to the target tone, (B) responses to the target word, (C) responses to the target image.
The responses are masked at p < 0.05, uncorrected. 0 sec refers to the onset of the respective
stimulus. Data is shown for all six subjects (S1-S6). The color boxes mark the frequency ranges
selected for the analyses of the responses to the different targets: target tone (8-15 Hz, purple),
target word (8-15 Hz, red), target image (15-25 Hz, blue). For the target tone a baseline from -

23
500 to 0 ms before target onset was used. For the target word and the target image a baseline
from -500 to 0 ms before the speech stream onset, which occurred approx. 560 ms before the
target word onset, was used.

Figure 3. Average and single subject CM-Pf responses to different stimulus events in the
oddball and multi-speaker tasks. (A) Neural responses to the target tone, deviant tone and
standard tone in the oddball task. (B) Neural responses to the target word and the two task words
in the multi-speaker task. 0 sec refers to the individual onset times of the different events. For the
stimulus events in (A) a baseline from -500 to 0 ms before the respective stimulus onsets was
used. For the stimulus events in (B) a baseline from -500 to 0 ms before the speech stream onset,
which occurred approx. 560 ms before the target word onset, was used. The target stimuli, which
required goal-oriented behavior selection, elicited the strongest response in CM-Pf.

Table 1 Summary of the clinical background for the individual patients. SCS: spinal cord
stimulation; MCS: motor cortex stimulation.

Case Age (at Cause of Pain Previous Medication at the Visual analogue
(sex) onset) neuropathic distribution surgery time of surgery scale: pain
pain for pain (min/max out of 10)

1(F) 57 (46) Nerve root Left leg SCS Tramadol 30 gtt 5/10
injury after 3x/d
spinal
surgery
2(M) 55 (26) Brachial Left arm None 7/10
plexus
injury
3(M) 55 (45) Nerve root Right leg SCS Pregabalin 600 mg 7/9
injury after 2x/d, Fentanyl 100
spinal μg/h
surgery
4(M) 45 (41) Wallenberg- Left face Pregabalin 225 mg 5/7
syndrome 2x/day
after

24
 brainstem
infarction
5(M) 25 (5) Resection of Right face MCS Gabapentin 600 4/8
parotid mg 2x/d,
gland tumor Pregabalin 75 mg
2x/d,
Novaminsulfon 1
g 2x/d,
Paracetamol 1 g
3x/d
6(F) 49 (47) Neck Right tongue Hydromorphon 8 4/7
dissection mg 6x/d,
for tongue Pregabalin 75 mg
tumor x2/d

Table 2. Stimulus material in the multi-speaker experiment

Target Action expression Task Word I Task Word II
Word

Goethe drücke gleich [press now] grün [green] zwei [two]

Los geht’s Schiller entscheide dich für [you choose] rot [red] und drei [three]
[Let’s go] Lessing wähle nun [choose now] weiss [weiss] [and] vier [four]
Heine nimm dismal [take this time] blau [blau] fünf [five]
Two spoken sentences were presented simultaneously and binaurally on each trial (see Fig. 1B),
one from a female and one from a male speaker. The target word, the action expression and the
two task words were always different between the two speakers. English translations for the
different expressions are provided in italic.

Table 3. Peak latencies of the neural responses (NR) to the different targets and reaction
times (RT) in ms
Subject ID Target Tone Target Word Target Image
NR RT NR NR RT

S1 184 438 130 340 1370

S2 164 393 - 550 1029
S3 193 533 300 - 2081
S4 164 402 - - 1760
S5 184 345 230 - 996
S6 193 422 230 - 750

25
26
27
Highlights
Recordings from human thalamic CM-Pf during auditory oddball and multi-speaker task
Neural responses in alpha/beta range to target tone, target word and target image
Greater responses to target tone and target word compared to other task events
Neural responses occur after target stimulus onset but before overt motor response