Intracranial recordings from patients with epilepsy detect electrical activity at multiple spatial and temporal scales to guide epilepsy surgery. Additionally, patients may participate in cognitive tasks that advance our understanding of high-level cognitive functions. In this symposium, we will first address iEEG oscillations in auditory cortex during the perception of auditory and visual speech (PM). Next, we will characterize the spatial scales of memory networks by simultaneously recording scalp EEG, hippocampal iEEG and single neuron activity (JS, FM). Finally, we show how intraoperative single pulse stimulation elicits perturbations at different temporal scales, which help to tailor the resection of the epileptogenic zone (GH). The symposium elucidates how multi-scale recordings serve to deal with scientific questions in neuroscience and clinical application.

Speech is essentially multisensory: the movements that we make when we speak are visible to our interlocutors. How do visual speech cues influence the processing of speech sounds by auditory cortex? To address this question, we showed short movies containing naturalistic speech to patients undergoing evaluation with intracranial EEG electrodes for epilepsy surgery. Auditory cortex responded to purely visual speech with an alignment of the phase of its low-frequency oscillations. Our results indicate that auditory cortex tracks the temporal dynamics of visual speech and support the role of neuronal oscillations in multisensory integration more generally.

Objective: Verbal working memory elicits workload-dependent theta and alpha oscillations in the frontal and parietal surface EEG (Michels, 2008), but the involvement of subcortical nodes is not known.

Methods: Epilepsy patients with electrodes in the hippocampus and on the scalp performed a modified Sternberg task with setsize 4, 6 and 8 letters. We analyzed the time-frequency profile and the phase locking value (PLV) in theta (4-8 Hz) and alpha (8-12 Hz) and high gamma (> 100 Hz) frequency bands while stimuli were encoded and retained in memory.

Results: In 9 of 10 participants, the theta/alpha PLV was elevated between hippocampus and scalp. In 4 participants, the PLV increased with setsize, predominantly towards the end of the retention period and to frontal and parietal electrodes. Two participants showed strong frontal midline theta and parietal alpha during retention, in agreement with our earlier scalp EEG study. Concurrently, the workload and/or the task conditions modulated the hippocampal theta/alpha and high-gamma power.

Conclusion: While hippocampal activity is known for visuospatial memory tasks, we show here hippocampal involvement also in a cortical network that is activated during verbal working memory and mediated by synchronized theta/alpha EEG oscillations.

Neurons in the human medial temporal lobe are tuned to semantic rather than perceptual features of visual stimuli. These neurons’ activity thus seems to correlate with contents of conscious experience, making them a prime candidate for a content-specific neuronal correlate of consciousness. In this talk, I will first present a study further exploring the nature of these semantic representations, and then present a series of studies aimed at exploring putative single neuron correlates of conscious awareness. One approach is to use adaptive algorithms to continuously estimate perceptual thresholds individually for patients and stimuli. Here, stimuli are presented either very near or clearly above or clearly below this threshold to disentangle effects of stimulus intensity from effects of awareness on spiking activity. These procedures were used to vary noise in visual stimuli when recording from medial temporal lobe neurons, and volume of auditory stimuli in one patient with electrodes in primary auditory areas. Another approach entailed using the attentional blink phenomenon to investigate spiking activity in response to stimuli that are sometimes seen and sometimes not detected at all. Here, we find that neurons fire in response to their preferred stimulus, even when participants reported absence of awareness of that stimulus. Remarkably, neuronal responses to unseen versus seen stimuli were delayed and temporally more dispersed, in addition to being generally attenuated in firing rate. Modulation of neuronal response timing and strength in response to seen versus unseen stimuli was found to increase along an anatomical gradient from posterior to anterior medial temporal lobe areas.

To be completed