Automatic vs. Manual Detection of High Frequency Oscillations in Intracranial Recordings From the Human Temporal Lobe.

Background: High frequency oscillations (HFOs) have attracted great interest among neuroscientists and epileptologists in recent years. Not only has their occurrence been linked to epileptogenesis, but also to physiologic processes, such as memory consolidation. There are at least two big challenges for HFO research. First, detection, when performed manually, is time consuming and prone to rater biases, but when performed automatically, it is biased by artifacts mimicking HFOs. Second, distinguishing physiologic from pathologic HFOs in patients with epilepsy is problematic. Here we automatically and manually detected HFOs in intracranial EEGs (iEEG) of patients with epilepsy, recorded during a visual memory task in order to assess the feasibility of the different detection approaches to identify task-related ripples, supporting the physiologic nature of HFOs in the temporal lobe. Methods: Ten patients with unclear seizure origin and bilaterally implanted macroelectrodes took part in a visual memory consolidation task. In addition to iEEG, scalp EEG, electrooculography (EOG), and facial electromyography (EMG) were recorded. iEEG channels contralateral to the suspected epileptogenic zone were inspected visually for HFOs. Furthermore, HFOs were marked automatically using an RMS detector and a Stockwell classifier. We compared the two detection approaches and assessed a possible link between task performance and HFO occurrence during encoding and retrieval trials. Results: HFO occurrence rates were significantly lower when events were marked manually. The automatic detection algorithm was greatly biased by filter-artifacts. Surprisingly, EOG artifacts as seen on scalp electrodes appeared to be linked to many HFOs in the iEEG. Occurrence rates could not be associated to memory performance, and we were not able to detect strictly defined "clear" ripples. Conclusion: Filtered graphoelements in the EEG are known to mimic HFOs and thus constitute a problem. So far, in invasive EEG recordings mostly technical artifacts and filtered epileptiform discharges have been considered as sources for these "false" HFOs. The data at hand suggests that even ocular artifacts might bias automatic detection in invasive recordings. Strict guidelines and standards for HFO detection are necessary in order to identify artifact-derived HFOs, especially in conditions when cognitive tasks might produce a high amount of artifacts.

Prediction of Cognitive Decline in Temporal Lobe Epilepsy and Mild Cognitive Impairment by EEG, MRI, and Neuropsychology.

Cognitive decline is a severe concern of patients with mild cognitive impairment. Also, in patients with temporal lobe epilepsy, memory problems are a frequently encountered problem with potential progression. On the background of a unifying hypothesis for cognitive decline, we merged knowledge from dementia and epilepsy research in order to identify biomarkers with a high predictive value for cognitive decline across and beyond these groups that can be fed into intelligent systems. We prospectively assessed patients with temporal lobe epilepsy (N = 9), mild cognitive impairment (N = 19), and subjective cognitive complaints (N = 4) and healthy controls (N = 18). All had structural cerebral MRI, EEG at rest and during declarative verbal memory performance, and a neuropsychological assessment which was repeated after 18 months. Cognitive decline was defined as significant change on neuropsychological subscales. We extracted volumetric and shape features from MRI and brain network measures from EEG and fed these features alongside a baseline testing in neuropsychology into a machine learning framework with feature subset selection and 5-fold cross validation. Out of 50 patients, 27 had a decline over time in executive functions, 23 in visual-verbal memory, 23 in divided attention, and 7 patients had an increase in depression scores. The best sensitivity/specificity for decline was 72%/82% for executive functions based on a feature combination from MRI volumetry and EEG partial coherence during recall of memories; 95%/74% for visual-verbal memory by combination of MRI-wavelet features and neuropsychology; 84%/76% for divided attention by combination of MRI-wavelet features and neuropsychology; and 81%/90% for increase of depression by combination of EEG partial directed coherence factor at rest and neuropsychology. Combining information from EEG, MRI, and neuropsychology in order to predict neuropsychological changes in a heterogeneous population could create a more general model of cognitive performance decline.

Reliability of EEG Measures of Interaction: A Paradigm Shift Is Needed to Fight the Reproducibility Crisis.

Measures of interaction (connectivity) of the EEG are at the forefront of current neuroscientific research. Unfortunately, test-retest reliability can be very low, depending on the measure and its estimation, the EEG-frequency of interest, the length of the signal, and the population under investigation. In addition, artifacts can hamper the continuity of the EEG signal, and in some clinical situations it is impractical to exclude artifacts. We aimed to examine factors that moderate test-retest reliability of measures of interaction. The study involved 40 patients with a range of neurological diseases and memory impairments (age median: 60; range 21-76; 40% female; 22 mild cognitive impairment, 5 subjective cognitive complaints, 13 temporal lobe epilepsy), and 20 healthy controls (age median: 61.5; range 23-74; 70% female). We calculated 14 measures of interaction based on the multivariate autoregressive model from two EEG-recordings separated by 2 weeks. We characterized test-retest reliability by correlating the measures between the two EEG-recordings for variations of data length, data discontinuity, artifact exclusion, model order, and frequency over all combinations of channels and all frequencies, individually for each subject, yielding a correlation coefficient for each participant. Excluding artifacts had strong effects on reliability of some measures, such as classical, real valued coherence (~0.1 before, ~0.9 after artifact exclusion). Full frequency directed transfer function was highly reliable and robust against artifacts. Variation of data length decreased reliability in relation to poor adjustment of model order and signal length. Variation of discontinuity had no effect, but reliabilities were different between model orders, frequency ranges, and patient groups depending on the measure. Pathology did not interact with variation of signal length or discontinuity. Our results emphasize the importance of documenting reliability, which may vary considerably between measures of interaction. We recommend careful selection of measures of interaction in accordance with the properties of the data. When only short data segments are available and when the signal length varies strongly across subjects after exclusion of artifacts, reliability becomes an issue. Finally, measures which show high reliability irrespective of the presence of artifacts could be extremely useful in clinical situations when exclusion of artifacts is impractical.

Electrical engram: how deep brain stimulation affects memory.

Deep brain stimulation (DBS) is a surgical procedure involving implantation of a pacemaker that sends electric impulses to specific brain regions. DBS has been applied in patients with Parkinson's disease, depression, and obsessive-compulsive disorder (among others), and more recently in patients with Alzheimer's disease to improve memory functions. Current DBS approaches are based on the concept that high-frequency stimulation inhibits or excites specific brain regions. However, because DBS entails the application of repetitive electrical stimuli, it primarily exerts an effect on extracellular field-potential oscillations similar to those recorded with electroencephalography. Here, we suggest a new perspective on how DBS may ameliorate memory dysfunction: it may enhance normal electrophysiological patterns underlying long-term memory processes within the medial temporal lobe.

Sleep-dependent directional coupling between human neocortex and hippocampus.

Complex interactions between neocortex and hippocampus are the neural basis of memory formation. Two-step theories of memory formation suggest that initial encoding of novel information depends on the induction of rapid plasticity within the hippocampus, and is followed by a second sleep-dependent step of memory consolidation. These theories predict information flow from the neocortex into the hippocampus during waking state and in the reverse direction during sleep. However, experimental evidence that interactions between hippocampus and neocortex have a predominant direction which reverses during sleep rely on cross-correlation analysis of data from animal experiments and yielded inconsistent results. Here, we investigated directional coupling in intracranial EEG data from human subjects using a phase-modeling approach which is well suited to reveal functional interdependencies in oscillatory data. In general, we observed that the anterior hippocampus predominantly drives nearby and remote brain regions. Surprisingly, however, the influence of neocortical regions on the hippocampus significantly increased during sleep as compared to waking state. These results question the standard model of hippocampal-neocortical interactions and suggest that sleep-dependent consolidation is accomplished by an active retrieval of hippocampal information by the neocortex.

3D source localization derived from subdural strip and grid electrodes: a simulation study.

OBJECTIVE: Little experience exists in the application of source reconstruction methods to recordings from subdural strip and grid electrodes. This study addressed the question, whether reliable and accurate 3D source localization is possible from the Electrocorticogram (ECoG). METHODS: The accuracy of source reconstruction was investigated by simulations and a case study. Simulated sources were used to compute potentials at the electrode positions derived from the MRI of a patient with subdural electrodes. Used procedures were the linear estimation (minimum norm) algorithm and the MUSIC (MUltiple SIgnal Classification) scan. RESULTS: Maxima of linear estimation were attracted to adjacent electrodes. Reliable localization with a localization error 15 mm was only achieved for about 35% of the original source positions. Maxima of the MUSIC metric were identical to original positions for simulations without noise. Noise reduced the percentage of reliable solutions down to a 79.0%. Electrode contacts distant to the source had small influence on localization accuracy. The case study supported simulation results. CONCLUSION: Reliable source reconstruction derived from ECoG can be achieved by the application of the MUSIC algorithm. Linear estimation needs additional compensation mechanisms. SIGNIFICANCE: MUSIC based 3D localization based on ECoG has the potential improving epilepsy diagnosis and cognitive research.

Sensory gating in intracranial recordings--the role of phase locking.

For patients with schizophrenia, a deficient gating (or filtering) of sensory input has been described. One major approach to study this sensory gating is to measure event-related potentials (ERPs) in response to paired clicks. In these experiments, sensory gating is quantified as amplitude reduction of the ERP components P50 and N100 from the 1st to the 2nd stimulus. In ERP studies brain electrical signals are averaged over single trials. Alterations in phase locking might be one factor contributing to the observed deficits in sensory gating, but findings have been inconclusive as yet. In particular, the contribution of different frequency bands to the deficit required further investigation. We studied N100 gating by intracranial recordings in a sample of epilepsy patients and subdivided the group into good and poor gators of the intracranial ERP component N100. Data were evaluated by frequency specific wavelet-based phase and power analyses. Poor N100 gators had an increased phase locking in the frequency range from 6.0-15.1 Hz after the 2nd stimulus, as compared to good gators. Other group differences were apparent already before the 2nd stimulus. Poor gators had less phase locked beta band activity (20.2-30.0 Hz) than good gators 200-315 ms after the onset of the 1st stimulus. Within the group of poor gators, lower values of phase locking in this frequency range were also associated with lower gating ratios. The reduced beta band response in response to the 1st stimulus may reflect poorer memory encoding of the 1st stimulus in poor gators. This in turn might lead to increased demands to process the 2nd stimulus.

Functional dissociations in top-down control dependent neural repetition priming.

Little is known about the neural mechanisms underlying top-down control of repetition priming. Here, we use functional brain imaging to investigate these mechanisms. Study and repetition tasks used a natural/man-made forced choice task. In the study phase subjects were required to respond to either pictures or words that were presented superimposed on each other. In the repetition phase only words were presented that were new, previously attended or ignored, or picture names that were derived from previously attended or ignored pictures. Relative to new words we found repetition priming for previously attended words. Previously ignored words showed a reduced priming effect, and there was no significant priming for pictures repeated as picture names. Brain imaging data showed that neural priming of words in the left prefrontal cortex (LIPFC) and left fusiform gyrus (LOTC) was affected by attention, semantic compatibility of superimposed stimuli during study and cross-modal priming. Neural priming reduced for words in the LIPFC and for words and pictures in the LOTC if stimuli were previously ignored. Previously ignored words that were semantically incompatible with a superimposed picture during study induce increased neural priming compared to semantically compatible ignored words (LIPFC) and decreased neural priming of previously attended pictures (LOTC). In summary, top-down control induces dissociable effects on neural priming by attention, cross-modal priming and semantic compatibility in a way that was not evident from behavioral results.

Word imageability affects the hippocampus in recognition memory.

Concrete words, whose meanings are readily imagined, are better remembered than abstract words. However, the neural correlates of this effect are poorly understood. Here, we investigated the effect of imageability on brain activity in the medial temporal lobe (MTL) processes underlying recognition memory. We recorded event-related potentials (ERPs) via depth electrodes from within the MTL in 14 patients with drug-resistant epilepsy. Patients performed a continuous word recognition task with words of high and low imageability (controlled for word frequency). Behaviorally, recognition performance was better for high, compared to low, imageable words. Two ERP components associated with recognition memory, the AMTL-N400 and the hippocampal late negative component, showed an old/new effect, but only the hippocampal P600 showed a main effect of imageability. We suggest that the hippocampal effect of imageability in recognition memory may be associated with conceptual or pictorial information processing of concrete words.

Using visual advance information: an event-related functional MRI study.

Our event-related functional MRI (efMRI) study investigates whether visual advance information (AI) affects rather perceptual or central response-related processing areas. Twelve subjects were required to make a go/no-go decision to a conjunction of a specific color and motion direction. The stimuli were preceded by a cue, providing 100% valid advance information about motion direction. Partial and full advance information (PAI and FAI) predicted possible targets, respectively, certain nontargets, neutral cues (NAI) gave no prediction. The time between cue and stimulus (stimulus onset asynchrony, SOA) was varied. A response benefit was found after PAI as compared with NAI. The benefit was small with a short SOA (150 ms), increased with intermediate SOA (450 ms) and sustained with long SOA (750 ms). Perceptual and central processing areas were more active with increasing SOA, but only central response-related processing areas were selectively modulated by cue information. In particular, supplementary motor area and bilateral inferior parietal lobe were more active with PAI than with NAI. If comparing NAI with FAI, more errors were made and activity was larger in central processing areas. Our results suggest that, depending on the processing time, cues providing perceptual information modulate central response-related processes.

Menstrual cycle-dependent neural plasticity in the adult human brain is hormone, task, and region specific.

In rodents, cyclically fluctuating levels of gonadal steroid hormones modulate neural plasticity by altering synaptic transmission and synaptogenesis. Alterations of mood and cognition observed during the menstrual cycle suggest that steroid-related plasticity also occurs in humans. Cycle phase-dependent differences in cognitive performance have almost exclusively been found in tasks probing lateralized neuronal domains, i.e., cognitive domains such as language, which are predominantly executed by one hemisphere. To search for neural correlates of hormonally mediated neural plasticity in humans, we thus conducted a functional magnetic resonance imaging study measuring brain activity related to a semantic decision task in the language domain. This was contrasted with a letter-matching task in the perceptual domain, in which we expected no steroid hormone-mediated effect. We investigated 12 young healthy women in a counterbalanced repeated-measure design during low-steroid menstruation and high-steroid midluteal phase. Steroid serum levels correlated with the volume and lateralization of particular brain activations related to the semantic task but not with brain activity related to the perceptual task. More specifically, bilateral superior temporal recruitment correlated positively with progesterone and medial superior frontal recruitment with both progesterone and estradiol serum levels, whereas activations in inferior and middle frontal cortex were unaffected by steroid levels. In contrast to these specific interactions, testosterone levels correlated nonselectively with overall activation levels by neural and/or vascular factor(s). In conclusion, our data demonstrate steroid hormone responsivity in the adult human brain by revealing neural plasticity in the language domain, which appears hormone, task, and region specific.

Covariation of spectral and nonlinear EEG measures with alpha biofeedback.

This study investigated how different spectral and nonlinear EEG measures covaried with alpha power during auditory alpha biofeedback training, performed by 13 healthy subjects. We found a significant positive correlation of alpha power with the largest Lyapunov-exponent, pointing to an increased dynamical instability of the EEG accompanying alpha enhancement. Alpha power amplification, moreover, was significantly correlated with a decrease of spectral entropy within the alpha range. This outcome reflects a sharpening of the alpha peak during biofeedback training. The fact that the sharpening effect clearly preceded the increase of alpha amplitude could be exploited in future biofeedback settings.

Is the nonREM-REM sleep cycle reset by forced awakenings from REM sleep?

In selective REM sleep deprivation (SRSD), the occurrence of stage REM is repeatedly interrupted by short awakenings. Typically, the interventions aggregate in clusters resembling the REM episodes in undisturbed sleep. This salient phenomenon can easily be explained if the nonREM-REM sleep process is continued during the periods of forced wakefulness. However, earlier studies have alternatively suggested that awakenings from sleep might rather discontinue and reset the ultradian process. Theoretically, the two explanations predict a different distribution of REM episode duration. We evaluated 117 SRSD treatment nights recorded from 14 depressive inpatients receiving low dosages of Trimipramine. The alarms were triggered by an automatic mechanism for the detection of REM sleep and had to be canceled by the subjects themselves. The REM episodes were determined as in undisturbed sleep-they had to include the remaining REM activity and were separated by 30 min without REM epochs. The frequency histogram of REM episodes declined exponentially with episode duration for each of the first four sleep cycles. The duration of nonREM intervals revealed bimodal distributions. These results were found consistent with the model assuming a reset of the ultradian cycle upon awakening. Whether REM or nonREM activity is resumed on return to sleep can be modeled by a random decision whereby the probability for REM sleep might depend on the momentary REM pressure.

Human scalp recorded sigma activity is modulated by slow EEG oscillations during deep sleep.

The EEG during deep sleep exhibits a distinct cortically generated slow oscillation of around and below 1 Hz which can be distinguished from other delta (0.5-3.5 Hz) activity. Intracranial studies showed that this slow oscillation triggers and groups cortical network firing. In the present study, we examined whether the phases of the slow oscillation during sleep stage 4 are correlated with the magnitude of sigma (12-16 Hz) and gamma (> 20 Hz) scalp activity. For this purpose, 10-min segments of uninterrupted stage 4 sleep EEG from 9 subjects were analyzed by applying wavelet techniques. We found that scalp recorded sigma, but not gamma, activity is modulated by the phases of the slow oscillation during deep sleep. Enhancement of sigma activity was observed to be triggered by the peak of the surface positive slow wave component, whereas reduction of sigma activity started around the peak of the negative component.

Human declarative memory formation: segregating rhinal and hippocampal contributions.

The medial temporal lobe (MTL) is the core structure of the declarative memory system, but which specific operation is performed by anatomically defined MTL substructures? One hypothesis proposes that the hippocampus carries out an exclusively mnemonic operation during declarative memory formation that is insensitive to content, whereas the rhinal cortex carries out an operation supporting memory formation indirectly. To explore the interaction between a salient item feature and memory formation, we contrasted neural correlates of memory formation of high- and low-frequency words. Event-related potentials (ERPs) were recorded via depth electrodes from within the MTL in nine epilepsy patients while they memorized single words. To assess memory formation, ERPs to words subsequently recalled in a free recall test were contrasted with ERPs to forgotten words. More high- than low-frequency words were remembered. High-frequency words led to distinct ERP subsequent memory effects in rhinal cortex and hippocampus. Low-frequency words, however, were only associated with the hippocampal ERP effect. The anatomically restricted interaction between word frequency and memory formation might indicate a semantically affected operation in the parahippocampal region supporting memory formation indirectly. By contrast, the missing interaction in hippocampal recordings might suggest a direct correlate of declarative memory formation that is insensitive to item properties.

GABA-ergic modulation of prefrontal spatio-temporal activation pattern during emotional processing: a combined fMRI/MEG study with placebo and lorazepam.

Various prefrontal cortical regions have been shown to be activated during emotional stimulation, whereas neurochemical mechanisms underlying emotional processing in the prefrontal cortex remain unclear. We therefore investigated the influence of the GABA-A potentiator lorazepam on prefrontal cortical emotional-motor spatio-temporal activation pattern in a combined functional magnetic resonance imaging/magnetoencephalography study. Lorazepam led to the reversal in orbito-frontal activation pattern, a shift of the early magnetic field dipole from the orbito-frontal to medial prefrontal cortex, and alterations in premotor/motor cortical function during negative and positive emotional stimulation. It is concluded that negative emotional processing in the orbito-frontal cortex may be modulated either directly or indirectly by GABA-A receptors. Such a modulation of orbito-frontal cortical emotional function by lorazepam has to be distinguished from its effects on cortical motor function as being independent from the kind of processing either emotional or nonemotional.

Suppression of EEG gamma activity may cause the attentional blink.

The attentional blink (AB) is an impairment of attention, which occurs when subjects have to report a target stimulus (T2) following a previous target (T1) with a short delay (up to 600 ms). Theories explaining the AB assume that processing of T2 is more vulnerable to decay or substitution, as long as attention is allocated to T1. Existing models of the AB, however, do not account for the fact that T2 detection accuracy reaches the minimum when T2 is presented after about 300 ms and not immediately following T1. Therefore, a new model is suggested, which is based on chronometrical considerations together with recent neurophysiological findings concerning the relation between the P3 event-related potential and the AB, the interaction between P3 and gamma oscillations, and the significance of the early evoked gamma band response. We hypothesize that suppression of the early gamma response to T2, accompanying the P3 related to T1, causes the AB.

The interaction of rhinal cortex and hippocampus in human declarative memory formation.

Human declarative memory formation crucially depends on processes within the medial temporal lobe (MTL). These processes can be monitored in real-time by recordings from depth electrodes implanted in the MTL of patients with epilepsy who undergo presurgical evaluation. In our studies, patients performed a word memorization task during depth EEG recording. Afterwards, the difference between event-related potentials (ERPs) corresponding to subsequently remembered versus forgotten words was analyzed. These kind of studies revealed that successful memory encoding is characterized by an early process generated by the rhinal cortex within 300 ms following stimulus onset. This rhinal process precedes a hippocampal process, which starts about 200 ms later. Further investigation revealed that the rhinal process seems to be a correlate of semantic preprocessing which supports memory formation, whereas the hippocampal process appears to be a correlate of an exclusively mnemonic operation. These studies yielded only indirect evidence for an interaction of rhinal cortex and hippocampus. Direct evidence for a memory related cooperation between both structures, however, has been found in a study analyzing so called gamma activity, EEG oscillations of around 40 Hz. This investigation showed that successful as opposed to unsuccessful memory formation is accompanied by an initial enhancement of rhinal-hippocampal phase synchronization, which is followed by a later desynchronization. Present knowledge about the function of phase synchronized gamma activity suggests that this phase coupling and decoupling initiates and later terminates communication between the two MTL structures. Phase synchronized rhinal-hippocampal gamma activity may, moreover, accomplish Hebbian synaptic modifications and thus provide an initial step of declarative memory formation on the synaptic level.