Increases in Medial Temporal Lobe BOLD Activity Co-Occur With Decreases in Theta Power During Usage of the Method of Loci
Material below summarizes the article Spatial Mnemonic Encoding: Theta Power Decreases and Medial Temporal Lobe BOLD Increases Co-Occur during the Usage of the Method of Loci, published on December 21, 2016, in eNeuro and authored by Marie-Christin Fellner, Gregor Volberg, Maria Wimber, Markus Goldhacker, Mark W. Greenlee, and Simon Hanslmayr.
How does the brain form new memories? Converging evidence shows that the key brain structures crucial for the formation of new memories are located in the medial temporal lobes (MTL). The same MTL regions are also part of the network of regions underlying spatial processing and navigation.
The idea that memory encoding and spatial processing rely on common resources is not new. The beneficial effect of spatial processing on memory encoding is known since ancient Greek times, when the Method of Loci, a mnemonic using spatial processing to enhance memory, was first described.
This mnemonic entails imagining to-be-remembered items along a familiar path. For example, someone imagines a grocery shopping list on their daily commute. Later, in the supermarket, all of the items on their shopping list can easily be retrieved by imagining the path again. To this day, memory athletes use the Method of Loci to memorize impressive amounts of seemingly arbitrary information. This overlap of efficient memory encoding and spatial processing renders the Method of Loci an interesting way to investigate MTL functions.
A question regarding memory processing in the MTL is how human MTL BOLD activity is related to brain oscillatory activity in the theta frequency range (approximately2-9Hz). In rodents, the most dominant and well-characterized oscillation in the MTL are theta oscillations. Increases in theta power in the rodent MTL have been reported during navigation and memory tasks, and the firing of place cells in the MTL is locked to the theta rhythm.
These findings have inspired several influential models that assign a crucial role to theta oscillations during spatial and mnemonic processing. Because of these links between theta oscillations and memory, most studies investigating the role of brain oscillation in human memory have focused on theta oscillations.
fMRI studies have consistently shown that MTL BOLD activity increases during memory formation. In contrast, studies investigating oscillatory correlates of successful memory formation have provided an inconclusive pattern of results: Some studies report increases of theta power during memory formation, whereas several others report theta power decreases during memory formation.
In the literature, it is commonly assumed that increases in theta power reflect involvement of the MTL in tasks. Therefore, the question whether increases or decreases of theta power characterize MTL activity during memory encoding remains unanswered.
In our study, we measured EEG and fMRI activity while human participants encoded lists of words using the Method of Loci and a non-spatial, associative semantic strategy: the Pegword method. Our goal was to investigate how MTL activity —measured by the fMRI BOLD signal — and brain oscillatory activity — recorded by scalp EEG — are related during spatial mnemonic encoding.
As the Method of Loci entails spatial processing and efficient memory encoding, we expected that the MTL should be especially active during this task and drive MTL related theta power changes.
Successful memory formation using the Method of Loci was unsurprisingly related to increases in MTL activity as measured by the fMRI BOLD signal. More surprising, however, was that the EEG data showed that successful memory formation is correlated with decreases in theta power. Items that were later successfully remembered elicited a stronger theta decrease than items that were later forgotten.
We also compared oscillatory power changes and BOLD activity during usage of the spatial Method of Loci with the non-spatial encoding strategy. Compared to this non-spatial, less efficient encoding task, the Method of Loci also led to stronger theta power decreases and stronger MTL BOLD activity increases. This pattern of concurrent theta decreases and BOLD activation increases during memory formation and spatial processing hints that decreases in theta power and increases in MTL activity might be related. A beamformer source reconstruction of the scalp EEG data also supported this assumption. According to the source reconstruction theta power, decreases were most likely generated in brain regions overlapping with the fMRI BOLD activity increases.
These findings suggest that MTL BOLD activity increases and theta power decreases indeed might be correlates of the same MTL-based neural processes. Decreases in theta power were most pronounced when the participants used the spatial Method of Loci. This mnemonic led to good memory performance and entails all cognitive processes traditionally linked to the MTL and the theta rhythm.
This negative relationship of theta power and MTL activity can be surprising because it is often implicitly assumed that increases — not decreases — in theta power might index the involvement of the MTL in a task. These findings challenge the long held view that theta oscillations in the MTL are positively related with fMRI responses.
The negative relationship between BOLD activity changes and theta power changes are, however, in line with prior reported negative correlations between low frequency power (below approximately 30 Hz) and BOLD activity changes in the neocortex. Consequently, decreases in low frequency power could serve as a general marker of neural activity.
From a mechanistic point of view, low frequency power decreases might be an important prerequisite for local neural information processing and enable the formation of fine-grained global long-range connection between brain areas.
Spatial Mnemonic Encoding: Theta Power Decreases and Medial Temporal Lobe BOLD Increases Co-Occur during the Usage of the Method of Loci. Marie-Christin Fellner, Gregor Volberg, Maria Wimber, Markus Goldhacker, Mark W. Greenlee, Simon Hanslmayr. eNeuro Dec 2016 DOI: 10.1523/ENEURO.0184-16.2016