Patterns of single-neuron activity during associative recognition memory in the human medial temporal lobe.

Electrophysiological activity in medial temporal lobe (MTL) structures is pivotal for declarative long-term memory. Single-neuron and microcircuit findings capitalizing on human microwire recordings from the medial temporal lobe are still fragmentary. In particular, it is an open question whether identical or different groups of neurons participate in different memory functions. Here, we investigated category-specific responses in the human MTL based on single-neuron recordings in presurgical epilepsy patients performing an associative long-term memory task. Additionally, auditory beat stimuli were presented during encoding and retrieval to modulate memory performance. We describe the proportion of neurons in amygdala, entorhinal cortex, hippocampus and parahippocampal cortex belonging to different response classes. These entail neurons coding stimulus-familiarity, neurons coding successful item memory, and neurons coding associated source memory, as well as the overlap between these classes. As major results we demonstrate that neurons responding to stimulus familiarity (old/new effect) can be identified in the MTL even when using previously known rather than entirely novel stimulus material (words). We observed a significant overlap between familiarity-related neurons and neurons coding item retrieval (remembered/forgotten effect). The largest fraction of familiarity-related neurons was found in the parahippocampal cortex, and a considerable fraction of all parahippocampal neurons was related to successful item retrieval. Neurons related to successful source retrieval were different from the neurons coding the associated information. Most importantly, there was no overlap between neurons coding item memory and those coding associated source memory strongly suggesting that these functions are facilitated by different sets of neurons.

Hyperfamiliarity for faces.

OBJECTIVE: To report 4 cases of hyperfamiliarity for faces (HFF) and review 5 previously reported cases. METHODS: We identified cases of HFF from PubMed search and references in prior reports. RESULTS: Three of our 4 cases had pathologic findings that were most extensive in the left temporal lobe. HFF occurred after a tonic-clonic seizure (cases 1 and 3), during simple partial seizures (case 2), and in the setting of an increase in simple partial seizure frequency but not during seizures (case 4). All 9 cases were adults with 1 or more seizures; symptoms first occurred after seizures in 5 cases and during seizures in 1 case. Ictal symptoms lasted from seconds to minutes and from 2 days to more than 7 years in the other 6 cases. The duration of HFF was not associated with the presence or extent of a structural lesion. While in several cases HFF appears to result from a postictal Todd paralysis, the mechanism underlying persistent cases is uncertain. CONCLUSIONS: This modality (visual)-specific and stimulus (face)-specific syndrome is associated with diverse structural, functional imaging, and neurophysiologic findings. Lesions are more often left-sided and involve the temporal lobe. Epilepsy and seizures were present in all 9 cases, suggesting a pathophysiologic relationship, which likely varies among cases. Although only reported in 9 patients, HFF is probably much more common than it is diagnosed.

Magnetoencephalographic correlates of different levels in subjective recognition memory.

In our current study we employed whole-head magnetoencephalography (MEG) to identify neurophysiological correlates (event-related fields, ERFs) of different phenomenologies in human recognition memory. Words which had previously been semantically processed were presented along with previously unstudied words. Via button presses, participants provided subjective indices of three forms of memory: confident recognition, familiarity-based recognition, and misclassification of previously presented items as new (no recognition, misses). Behavioral results revealed a clear distinction between confident recognition (shortest reaction times) and familiarity-based recognition and misses, respectively, and physiological data pointed to bilateral anterior and left anterior/central regions in which magnetic field patterns were directly related to word recognition from approximately 300 ms to 500 ms after word onset. In the context of the prevalent dual process controversy on the roles of familiarity and recollection in recognition memory, we first highlight that two operationalizations of recollection need to be differentiated: We argue that a strategic search for a particular contextual feature stands in clear contrast to the fast and incidental availability of some contextual feature and derive experimental and behavioral indicators for either form of recollection. These indicators are used to select from manifold cognitive neuroscientific work on recognition memory in order to further discuss the neurocognitive characteristics of incidental recollection in contrast to other forms of episodic memory.

Neurocognitive correlates of incidental verbal memory encoding: a magnetoencephalographic (MEG) study.

In our current study, we applied whole head magnetoencephalography (MEG) in a subsequent memory paradigm. Magnetic fields were recorded while 20 healthy subjects (10 females, 10 males) incidentally encoded words during semantic and structural verbal processing tasks. Physiological data were then sorted according to the performance in subsequent memory tests and according to levels of processing, respectively, and analyzed for gender effects. Behavioral results show a clear advantage of semantic processing over structural processing with respect to retrieval success for both females and males. Despite alikeness of behavioral data, MEG results show considerable differences between males and females concerning both subsequent memory effects and levels of processing effects. For male subjects, we identified more distinct magnetic fields in anterior regions for subsequently remembered than for subsequently forgotten words (latency range from 300-650 ms after word onset) and for semantic processing than for structural processing, respectively. For female subjects, corresponding magnetic field differences pointed to posterior regions (subsequent memory effects from 450 ms to 750 ms after word onset). No qualitative differences were observed between subsequent memory effects during semantic processing compared to subsequent memory effects during structural processing. We try to reconcile results from male data with previous findings concerning subsequent memory effects by proposing the concept of width of processing, which holds that incidental memory formation is mediated by frontal activity on a physiological level and brought forward on a cognitive level by enhanced associating imposed by the task demands of semantic processing. Female data cannot be fully incorporated in this framework, but all the more prompt further gender-specific analyses of subsequent memory effects.