Experimental Brain Research

, Volume 156, Issue 4, pp 513–517

Attention to pain is processed at multiple cortical sites in man


  • Shinji Ohara
    • Department of NeurosurgeryJohns Hopkins Hospital
  • Nathan E. Crone
    • Department of NeurologyJohns Hopkins Hospital
  • Nirit Weiss
    • Department of NeurosurgeryJohns Hopkins Hospital
  • Hagen Vogel
    • Institute of Physiology and PathophysiologyJohannes Gutenberg University
  • Rolf-Detlef Treede
    • Institute of Physiology and PathophysiologyJohannes Gutenberg University
    • Department of NeurosurgeryJohns Hopkins Hospital
    • Department of NeurosurgeryJohns Hopkins Hospital
Research Note

DOI: 10.1007/s00221-004-1885-2

Cite this article as:
Ohara, S., Crone, N.E., Weiss, N. et al. Exp Brain Res (2004) 156: 513. doi:10.1007/s00221-004-1885-2


Painful cutaneous laser stimuli evoked potentials (LEPs) were recorded over the primary somatosensory (SI), parasylvian, and medial frontal (MF) cortex areas in a patient with subdural electrode grids located over these areas for surgical treatment of epilepsy. The amplitudes of the negative (N2*) and positive (P2**) LEP peaks over SI, parasylvian, and MF cortex were enhanced by attention to (counting stimuli), in comparison with distraction from the stimulus (reading for comprehension). Late positive deflections following the P2** peak (late potential—LP) were recorded over MF and from the lateral premotor regions during attention but not during distraction. These findings suggest that attention gates both early (N2*) and late (P2**) pain-related input to SI, parasylvian, and MF cortical regions while the later components (LP) are specifically related to attention.


Attention to painDistributed cortical networkLaser evoked potentialsEpilepsySubdural electrode gridsSomatosensory cortex


Directed attention can powerfully modulate pain perception. Under experimental conditions, the analgesic effect of distraction may equal that of narcotic analgesics (Scharein et al. 1998). The extent to which attention modulates early stages of pain processing cannot be addressed by positron emission tomography (PET) or functional MRI because these techniques have limited temporal resolution.

Attention-related modulation of early stages of pain processing can be studied through potentials evoked by painful cutaneous laser stimulation (LEPs). The major peaks of LEPs, N2, P2, and P3, are named by the polarity of the maximum recorded at the scalp midline, whereas the earliest negativity N1 is maximal over parasylvian cortex. Modeling of generators of the scalp EEG (Schlereth et al. 2003; Tarkka et al. 1993; Valeriani et al. 1996), MEG (Kakigi et al. 1995; Kanda et al. 2000; Ploner et al. 1999), depth recording (Frot et al. 2001), and direct cortical recording (Kanda et al. 2000; Lenz et al. 1998a, 1998b, 2000b) with source analysis (Kanda et al. 2000; Vogel et al. 2003) has revealed that these peaks are generated by current sources in the medial frontal (MF, including anterior cingulate—ACC, supplementary motor area—SMA), primary somatosensory (SI), and parasylvian cortices.

Whereas LEP P2 and P3 components measured from the scalp are modulated by endogenous factors such as attention, early stage modulation of LEPs (N1, N2) is controversial (Beydoun et al. 1993; Garcia-Larrea et al. 1997; Legrain et al. 2002; Siedenberg et al. 1996). The cortical location of sites where LEPs are modulated by attention has been estimated by source analysis of scalp recordings (Schlereth et al. 2003), which show significant modulation in parasylvian cortex, but not in SI.

We now describe the cortical areas where attention-related modulation of LEPs occurs, based on recordings directly from the cortex, for the first time. Subdural recordings were made from SI, parasylvian, and MF regions in a patient with subdural grid electrodes implanted for investigation of epilepsy. The results demonstrate simultaneous early and late stage attention-related modulation of LEPs in all three cortical areas.

Subjects and methods

The subject was a 21-year-old woman with medically intractable seizures since age 10. Neurological examinations and brain magnetic resonance images were normal. Subdural electrode grids were placed over the frontal-central-parasylvian cortex and the medial wall of the left hemisphere (Fig. 1). The protocol for these studies was approved by the Institutional Review Board of the Johns Hopkins University. The patient signed an informed consent for the studies.
Fig. 1

Distribution of the negative peak of the laser evoked subdural potential during attention condition and representative waveform N2* peaks was recorded at locations localized by cortical maps and atlas coordinates (x, y, z—see Talairach et al. 1988) in primary somatosensory (SI: 45, −28, 53), parasylvian (PS: 60, −6, −10) and medial frontal (MF: −1, −8, 31) cortex regions. Schematic maps show the distribution of N2* amplitude during the attention condition. Note the clear N2* amplitude difference between attention (counting) and distraction (reading) conditions (CS central sulcus, CiS cingulate sulcus, MCiS marginal branch of CiS, SF sylvian fissure)

The grid consisted of platinum-iridium circular electrodes (2.3 mm diameter) with a center-to-center distance between electrodes of 1 cm (Ad-Tech, Racine, WI). The signals from 89 electrodes were amplified and band-pass filtered at 0.1–300 Hz (Astro-Med, Inc., West Warwick, RI), and referenced to a single subdural electrode chosen for its relative inactivity and maximum distance from the active electrodes (LAF16, near the frontal pole). The signals were digitized at 1,000 Hz and recorded on a hard disk along with stimulus markers. Cortical function was mapped by cortical stimulation and by recording somatosensory evoked potentials (SEPs). Cortical stimulation was carried out to localize sensory, motor and language areas (Lesser et al. 1994). SEPs were recorded by electrically stimulating the median nerve at the wrist contralateral to the grid.

During this protocol the patient wore goggles and reclined on a bed, quietly wakeful with eyes open. Cutaneous heat stimulation was delivered to the right hand dorsum (contralateral to the grid) by a Thulium YAG laser (Neurotest, Wavelight Inc., Starnberg, Germany). The duration of each pulse was 1 ms and the beam diameter was 6 mm. To avoid fatigue or sensitization, the laser beam was moved at random to a slightly different position for each pulse. A total of 38 laser pulses at 720 mJ were delivered randomly with interval of 8–10 s for each recording run.

We employed two runs of stimuli for each of two conditions characterized by attention to, and distraction from, the stimulus. During the attention runs, the subject was asked to count the number of painful stimuli, and to report the total/run. During distraction runs, the subject was asked to read a magazine article for comprehension, as demonstrated by the response to questions about the article. The sequence of runs was “attention, distraction, distraction, attention.”

The positions of subdural electrodes were determined relative to the central sulcus (CS), the sylvian fissure, and the cingulate sulcus as located by intraoperative photographs and drawings as in previous studies (Crone et al. 1998; Lenz et al. 1998a, 1998b, 2000). Databases of the three-dimensional (3D) postoperative computed tomographic scan (CT) and the 3D preoperative MRI scan were also coregistered.

Multi-channel signals were re-montaged using an average reference to minimize the influence of location and activity of the reference electrode (Crone et al. 1998). The signals were averaged time-locked to the onset of laser with a time window of 1.0 s including the 0.1 s pre-stimulus period. Trials with artifacts or large baseline fluctuation were excluded by visual inspection. Averaged waveforms were obtained after confirming the reproducibility of results between two recording runs (approximately 35 trials each) for each condition.

Peak latencies and amplitudes were measured from the averaged waveforms and peak amplitudes were measured from the baseline value, defined as the average during pre-stimulus period. Peak latencies were measured at the time of the peak amplitude and peaks were regarded as significant when the peak amplitude was above 2 SD of the baseline value for this electrode. In describing the latencies of the LEPs, we referred to the largest mostly negative wave over MF (Fig. 1, MF) as N2*, and to the subsequent large, positive wave over MF (Fig. 1, MF) as P2**. The asterisks indicate that the latencies of the subdural potentials are somewhat different from reports of scalp recordings (Ohara et al. 2003). The late potential (LP) was based on the recordings from MF cortex (Fig. 2, MF). The effect of attention on LEP amplitudes was assessed by taking a ratio of absolute amplitude during attention (AMPa) and distraction (AMPd) conditions. Specifically, the % amplitude change, defined as 100–100x(AMPd/AMPa) (%), was obtained for each electrode for each peak. The % amplitude change in each region for each peak was tested to establish whether it was significantly different from 0 (t-test). We compared the change in % amplitude between three regions with one-way ANOVA for each peak (N2* or P2**). Post hoc analysis was performed with Tukey’s Honestly Significant Difference (HSD) Test.
Fig. 2

Distribution of positive P2** (A) and LP (B) peaks of the laser evoked subdural potential during attention condition and representative waveforms. A P2** peaks were recorded from primary somatosensory (SI: 37, −25, 59), parasylvian (PS: 63, −12, −6) and medial frontal (MF: −1, −16, 33) cortex regions. The amplitude of the P2** peak was strongly enhanced during the attention task. B LP was recorded from the MF region and a part of the lateral premotor area (38, −11, 48) only during the attention condition (counting the laser stimuli). Conventions as in Fig. 1


In both attention condition runs, the patient counted the number of laser pulses correctly (n=38). After the two distraction condition runs, the patient correctly answered the questions about the text that she had read (one question for each run). Average pain rating for the laser stimulus was 5/10 on a scale from 0 (no pain) to 10 (the most intense pain imaginable) during the attention condition and 1/10 for the second distraction series. The patient felt pain only 2 or 3 times during the first distraction series.

Significant LEP peaks were recorded from three distinct regions: near the hand area of SI, the parasylvian, and the MF regions. In each of the three regions, an initial negativity at 133–157 ms peak latency (N2*, Fig. 1) was followed by a positivity at 206–216 ms peak latency (P2**, Fig. 2A). The effect of distraction on LEP amplitudes was a significant decrease in all three regions for both the N2* and P2** peaks, excepting the P2** peak recorded over the MF area (Table 1). These regional differences were reflected in the ANOVA results, which showed a significant difference between regions for P2** amplitude (F(2,28)=4.2, p=0.02), but not N2* (F(2,17)=0.6, p=0.56). The lack of significant modulation of P2** in MF by attention is caused by the substantial variance across electrodes (decreases as shown in Fig. 2, but increases at other electrodes), consistent with imaging studies (Derbyshire et al. 1998).
Table 1

Mean LEP peak latency and absolute amplitude for primary somatosensory (SI), parasylvian and medial frontal (MF) cortex regions (mean ± SEM). Mean latency was calculated only for significant peaks. Mean amplitude was calculated for both significant and non-significant peaks, assuming that the amplitude of non-significant peak was zero. Number of electrodes is given in brackets









Latency (ms)

137±2 (11)

157±2 (6)

206±3 (13)

220, 225 (2)

Amplitude (μV)

43±7 (11)

14±4 (11)

35±2 (13)

3±5 (13)

% amp. change




Latency (ms)

133±2 (4)

157 (1)

216±6 (9)

250, 250 (2)

Amplitude (μV)

27±13 (4)

8±8 (4)

68±11 (9)

4±3 (9)

% amp. change




Latency (ms)

154±4 (5)

143, 162 (2)

206±5 (9)

223±10 (5)

Amplitude (μV)

60±9 (5)

8±5 (5)

42±8 (9)

19±6 (9)

% amp. change



**p<0.00 and *p<0.01, significant amplitude change between attention and distraction tasks (t-test)

We also recorded large late positive deflections (LP) with the peak latency at approximately 350 ms only during the attention condition and only over the MF region and the lateral premotor area (Brodmann’s area 6, Fig. 2B, upper). These characteristics suggest that the LP is likely to represent a laser-evoked P3 potential of the type associated with the occurrence of infrequent (novel) events (Kanda et al. 1996; Siedenberg et al. 1996). Over the MF region the location of the maximum of the P2** (Fig. 2A, lower) is similar to that of N2* (Fig. 1, lower), and both are caudal to that of the LP (Fig. 2B, lower).


By using simultaneous subdural LEP recordings directly over SI, parasylvian and MF cortical areas, we have demonstrated both early (N2*) and late (P2**, LP) modulation of cortical activity by attention to painful laser pulses. These early and late stages may be related to sensory-discriminative (early) and affective-motivational or cognitive-evaluative (late) components of pain perception (Price 1988, 2000). The prolonged, widespread effects of attention suggest that attention-related modulation of pain perception is a broadly distributed cortical process.

Previous studies reported attentional modulation of LEPs recorded from individual (vertex and parietal) leads which could not distinguish different cortical areas (Beydoun et al. 1993; Garcia-Larrea et al. 1997; Zaslansky et al. 1996). These studies showed that the amplitudes of the LEP components N2 (220–240 ms) and P2 (330–370 ms) were enhanced by attention or attenuated by distraction (Beydoun et al. 1993; Garcia-Larrea et al. 1997; Zaslansky et al. 1996). Early N1 peaks (120–200 ms) were modulated by spatial attention (Legrain et al. 2002; Schlereth et al. 2003) while late positive peaks corresponding to P3 (300–600 ms) were increased by conscious detection of infrequent events (novelty), and by the attention condition of the present study.

Our results recorded directly from the cortical surface in humans demonstrate that attention-related changes in N2* and P2** are significantly different from zero except for the P2** component recorded over MF. This exception was unexpected, because scalp recorded P2 was robustly modulated in previous studies, and ACC sources have a major contribution to this component (Chen et al. 1995; Tarkka et al. 1993; Valeriani et al. 1996). The electrode nearest to the mean dipole location of the cingulate generator in previous studies (Garcia-Larrea et al. 2003) indeed showed a large signal increase during attention (Fig. 2A), but other electrodes exhibited little or even opposite modulation. Hence, the variability in the effect of attention on LEPs in MF may be consistent with the suggestion that the ACC contains multiple functional areas, which are subject to interindividual differences (Derbyshire et al. 1998). A source analysis of scalp LEPs suggested that attention-related increases only occurred for the early negativity in parasylvian cortex, and the late positivity in MF cortex (Schlereth et al. 2003), perhaps due to limited sensitivity of the method.

Studies of medullary dorsal horn, and thalamus in trained monkeys, have reported attentional task-related modulation of the single neuron responses to noxious stimuli (Bushnell et al. 1984, 1989, 1993). Task-related modulation was reported for neurons in the medial thalamic nuclei, possibly projecting to MF cortex (Bentivoglio et al. 1993), but not for lateral nuclei projecting to SI (Bushnell et al. 1993). Thus the task-related modulation of the earliest (N2*) components in some cortical areas of this study may reflect activity afferent from spinal and thalamic centers signaling painful stimuli and projecting to cortex (Price 2000; Willis 1985).

The present results also demonstrate that the LP occurred in the latency range of the LEP peak associated with novel stimuli (Legrain et al. 2003; Lenz et al. 2000; Zaslansky et al. 1996), during the attention condition, over ACC, SMA and lateral premotor areas (Brodmann’s area 6). Studies using the same subdural-recording technique demonstrate that these regions are involved in movement preparation and in short-term memory of sensory stimuli which cue subsequent motor tasks (Ikeda et al. 1999; Matsumoto et al. 2003). Thus, these cortical areas may be involved in the motivational aspect of voluntary movements occurring in response to painful stimuli.


This work was supported by the National Institutes of Health-National Institute of Neurological Disorders and Stroke (NS38493 and NS40059 to FAL) and Deutsche Forschungsgemeinschaft (Tr236/13-3 to RDT). We thank D. Jackson and L.H. Rowland for excellent technical assistance.

Copyright information

© Springer-Verlag 2004