Functional magnetic resonance imaging (fMRI) has developed into a tool that is extensively used in non-invasive brain imaging. It provides information about cerebro-vascular activity throughout the whole brain with excellent spatial localisation, yet it is limited by the poor temporal resolution it offers, which is in the order of seconds. On the other hand, electroencephalogram (EEG) is recorded directly from the scalp of the subject and can provide information about neurophysiological activity with a very precise temporal resolution, in the order of milliseconds. The disadvantage of EEG however, is that localisation of the source of electrical activity within the brain is quite difficult.
Therefore the integration of these two techniques is highly desirable, and would allow the exploitation of the advantages of both techniques – high spatial resolution of fMRI and high temporal resolution of EEG. The technique would be widely applicable in all areas of neurophysiological research, but particularly in pain studies, where combined use of the two techniques could lead to the identification of distinct patterns of brain activity indicative of pain and pro-nociceptive sensitisation in healthy subjects and chronic pain patients. These patterns could prove useful in the assessment of the analgesic efficacy of novel analgesic compounds, adding to the desirability of co-registration of pain evoked potentials (such as those stimulated by contact heat) with EEG and fMRI.
Objective markers of pain could be less variable and/or more sensitive to analgesic treatments (i.e. EPs, fMRI). The present study enables the use of these markers to simultaneously assess pharmacodynamic-pharmacokinetic relationships and antihyperalgesic activity in single dose studies in experimental pain models in humans, and in pain patients (e.g. of novel agents that block TRPV1, the heat and capsaicin receptor). Pain biomarkers are needed to provide early pivotal information on efficacy, dose-response and time-course of TRPV1 antagonists, for strategic rapid and cost-effective drug development. Further, neuropathic pain patients would be expected to have reduced pain-evoked potential amplitudes but increased fMRI activation, different from volunteer or inflammatory models/conditions.
The use of simultaneous EEG and fMRI in pharmacological studies of novel analgesics would allow (i) measurement of the combined response to the same pharmacological intervention when the plasma concentration of the analgesic is at its peak (i.e. at the same time-course); (ii) recordings can be analysed at the single subject level, reducing inter-session variability, which is well recognised with regard to pain scores, and (iii) selection of the parameter which enables possible reduction in dose in clinical trials.
Simultaneous EEG and fMRI has already been utilised in studies of epilepsy [1
], sleep [2
], studies of human alpha activity [3
] and the study of auditory [5
], visual [7
], and motor activity [9
]. In the field of pain research, laser and electrically stimulated pain evoked potentials have been successfully co-registered using EEG and fMRI [10
], but CHEPS has the advantages of a larger area of stimulation, the ability to apply repetitive stimuli to the same cutaneous area without inducing erythema, and simplicity of use in the clinic (i.e. no eye protection required, easy to move location); however, one disadvantage of CHEPS is that in order to avoid habituation the thermode must be physically moved between stimuli.
Simultaneous EEG/fMRI is a readily feasible application yet the MR environment poses a number of technical challenges, limitations and safety issues (for a review of technical and safety issues see [12
]). The gradient switching fields and radiofrequency (RF) pulses of the MR scanner can create currents in conducting loops that can potentially cause heating of the electrodes and burns at the point of contact with the scalp of the subject. Therefore it is recommended to avoid loops and crossing over of electrode wires, and introduce current limiting resistors to electrode wires.
The gradient fields and RF pulses of the MR scanner can lead to artefacts that can obscure the signal recorded from the EEG, as can movements within the static magnetic field; movement of the head, talking, swallowing and even the pulsatile motion of the heart (ballistocardiogram artefact). The magnet's helium cold pump and gradient coils also produce mechanical vibrations that can be picked up by the EEG wires in the static magnetic field and converted to electrical noise. All these issues need to be addressed and overcome during experimental set-up and processing of EEG and fMRI data, in terms of patient safety and also preserving the quality of the data collected.
Until recently, quantifiable contact heat evoked potentials in EEG had been hard to elicit, due to technical limitations i.e. slow temperature rise and fall times. However, the contact heat evoked potential stimulator (CHEPS) has been designed with a maximum (adjustable) temperature rise time of 70°C/s. CHEPS can stimulate the differential thermal thresholds of receptors innervated by Aδ and C nociceptive nerve fibres, and has been shown to selectively excite these fibre subtypes in human hairy and glabrous skin [14
]. The latencies and amplitudes of heat evoked Aδ potentials stimulated by CHEPS have been shown to be robust and reproducible [15
], despite the disadvantage posed by the slow temperature rise time of the heat stimuli produced by CHEPS (70°C/s compared to a rise time of greater than 1000°C/s reported with lasers) which may lead to a reduction in temporal summation and thus a less synchronous afferent volley [17
The compatibility of CHEPS with fMRI further widens its scope of application both in healthy volunteers for research purposes and in chronic pain patients. Reproducibility of blood oxygen level dependent (BOLD) responses and pain scores to CHEPS stimulation have already been illustrated in healthy volunteers [18
In this human volunteer study we have assessed the feasibility of monitoring contact heat evoked Aδ potentials with simultaneous EEG and fMRI, and determined their relationship to the ratings of evoked pain.