TY - RPRT T1 - Theta’s functional role in encoding, nocturnal reactivation and retrieval for the formation of declarative memory traces AU - Paßmann, Sven AU - Rasch, Björn DO - 10.17605/OSF.IO/YSW5K UR - https://osf.io/ysw5k/ AB - The formation of memory traces is a two-stage process. Newly encoded items, collected during wakefulness and stored temporarily into the Hippocampus, must be redistributed to a long- term memory and integrated into pre-existing memory traces due to their instability and vulnerability to subsequent encoded information [1]. The success of the so-called memory consolidation depends on how the information were collected and on their underlying oscillatory activities during wake and sleep. In short, memory research differentiates three processes in information processing for declarative memory items: encoding, consolidation and retrieval. Learning (encoding) of items can be done explicitly (intentional or incidental) or implicitly. Explicit encoding is more robust compared with implicit encoding. Intentionally learned items are remembered better than incidentally learned ones (see review [2]). The subsequent process called memory consolidation gradually transforms these newly encoded, hippocampal-stored, labile information into stable traces and integrates them into pre-existing memory. This process has been described by Diekelmann and Born with the Active Systems Consolidation Theory [3], postulating a synchronized interaction between nocturnal slow oscillatory activity (SO), hippocampal sharp wave ripples and thalamo-cortical spindles. The retrieval of information is the process to gain access to previously consolidated information. Here, memory traces are transformed once again from a stable to a labile state [1,4,5]. During wakefulness higher frequency bands such as beta and gamma, but also alpha and theta dominate the oscillatory activities. Here, theta oscillatory activity (generated in the Hippocampus [6] and in neocortical structures [7] ranging between 4 and 7.5 Hz) shows an opposite behaviour to neocortically generated alpha activity indicating a functional coupling of both oscillatory activities [8]. Especially, the increase of theta oscillatory activity is discussed as enabling long term potentiation (LTP) [9]. Previous findings revealed that LTP-like processes can be best induced during theta [10], primarily during the positive phase [11], whereby the strength of the induced LTP is highly correlated with the increase in theta power [12]. Additionally, theta oscillations are also assumed to play a critical role in binding spatial information which is represented by oscillations in the gamma range [13,14]. In detail, when hippocampal place cells (representing a distinct spatial information) fire during a particular phase of a theta cycle, other place cells are likely to fire afterwards forming a temporal sequence of information to a neural code. Similar to wakefulness, nocturnal theta activity shows the highest power values during sleep stage 1 alongside with alpha activity. With increasing sleep depth theta and alpha activity decreases until non-REM sleep stage 3 [15] which reflects the process S of the two-process model of sleep regulation by Borbély [16]. Additionally, hippocampal-generated theta activity [7] is also a fundamental electrophysiological feature of the REM sleep stage [17], which is considered to be involved in consolidating spatial, emotional and procedural information [18] as well as in integrating declarative memories into pre-existing traces [3]. Interestingly, hippocampal theta oscillations during transition from sleep to wakefulness appear alongside neocortical theta activity. They also last significantly longer and are slightly faster than those in the REM sleep stage [7]. More and more evidence shows that the oscillatory activity in the brain during encoding, reactivation and retrieval plays a critical role for information processing. Numerous studies have consistently shown that higher theta activity (4-7 Hz) before [19–21] and during encoding [9,22– 25] is linked to a better retrieval performance at a later time point, independently of presenting pictures or words - often accompanied by a decrease in alpha desynchronization (8-13 Hz). The so-called subsequent memory effect (SME; [26]) was found in several study designs: before encoding in a free word recall paradigm [27], but also in word [21,28,29] and picture recognition [19,20,25]. During encoding SME was found for cued recall [23,30] as well as free recall [31], but also for intentional and for incidental learning [19,20,24,30,31]. These results are supported by studies which showed that higher resting state theta power in healthy older adults is also correlated with better cognitive performance [32]. This could not yet been shown for young adults. In addition, higher theta activity does not predict later memory performance in general, but shows a context dependency of the consolidated items for later retrieval [33]. Here, a successful retrieval of consolidated items depends on the surrounding context which must not differ between encoding and later retrieval. The result might contribute to the idea of theta activity as a tagging feature: theta activity in combination with prefrontal networks seems to support anticipatory aspects of behaviour indicating a tagging of relevant information for later retrieval [2]. Furthermore, previously consolidated memory traces are reactivated during sleep and retrieval [1], changing from a stable to a labile state once again. Given that items are also labile during encoding, it is plausible to assume that theta activity could play a similar role during reactivation in sleep and later retrieval. The theta-related context dependency mentioned above and the role of theta activity as a coordinating activity between hippocampus and prefrontal cortex [34] lets assume that theta activity is necessary to track relations among items and contextual features [35] also during retrieval [36]. This is supported by several studies which indicate a significant impact of theta activity in recognition [37–39], showing higher theta activity during retrieval processes which might reflect the attempt to retrieve. The authors discussed these findings as process specific, not primarily related to the actual access and retrieval of memory traces [38]. This assumption was made because they did not find any difference between hits and correct rejections. Additionally, a study by Schreiner and colleagues [40] revealed that nocturnal theta activity also seems to play a role for successful consolidation of nocturnally reactivated memory traces. Here, the authors could show that higher theta activity was associated with a better memory performance of those words which were cued during sleep. Their study revealed a correlation of gain in word pair recall after sleep with an increase of theta power in right frontal and left parietal electrode positions during cueing. A possible explanation for this result could be given by a study of Benchenane et al. [41]. Here, it was shown that prefrontal cell assemblies which fire during encoding (facing a theta-related coherence in the prefrontal-hippocampal circuit), show an increase in probability for reactivation during subsequent slow wave sleep (SWS). Interestingly, during sleep a context-depending reactivation is also fundamental for a successful reactivation [42] and therefore consolidation of declarative memory items [43]. During sleep theta activity is still present as a coordinating activity between hippocampus and neocortex [7] and could serve here as a specific feature necessary to bind context-dependent items which are transformed into labile state during reactivation. Note, that most of these studies were based on visually presented items (see review [2]). Results by Gaab et al. [44] and others [45–47] implicate that auditory learning also benefits from subsequent sleep. Conway and colleagues [48] could show that a modality preference exists for auditory presentation compared to visual or tactile presentation. Moreover, recent studies focusing on the underlying oscillatory activity (event-related potentials) during encoding found an SME for items presented acoustically [28,29,49] as well as an appropriate increment in nocturnal theta activity when items were reactivated during SWS [40]. In sum, theta oscillations are supposed to participate in memory encoding, consolidation and retrieval processes. Alongside findings that theta activity works as a binding factor [13,14], it also seems to be fundamental activity for successfully encoding and retrieving items. But as most of the cited studies have only used correlational approaches to examine the role of theta activity for memory, the crucial question remains whether this oscillation plays a functional role in memory. Transcranial direct current stimulation (tDCS) is a highly promising method for examining the functional role of brain oscillations. Nitsche et al [50] published one of the first studies with electrical stimulation of human brains through the skull. They explained the results as a direct consequence of neuronal depolarisation as it was found in animal studies [51,52]. The increase of neuronal excitability leads to LTP-like processes [53,54] which are one of the main factors in the consolidation process [3]. Support for this idea came from studies which used calcium and sodium channel blockers which reduced or even abolished stimulation effects [55]. In comparison to tDCS, transcranial alternating current stimulation (tACS) is a relatively new approach in modulating oscillatory activity. Similar to oscillatory tDCS the main effect of tACS is to modulate and entrain the ongoing rhythmic brain activity, but without LTP-like effects [56]. However, although slight neuroplastic excitability modifications have been already described [57,58], it is more frequently used as a tool to examine the functional role of distinct brain patterns which underlie cognitive functions [59]. The systematic examination of theta and its functional role in memory processes, especially in encoding, started with studies [60–62] which examined the impact of theta-based stimulation. They revealed a positive influence of the stimulation on working memory compared to a sham- condition. With regard to declarative memory, a recent study [63] was able to show that theta- tACS applied during encoding improved the retrieval performance in a picture-word training paradigm compared to sham. Unfortunately, the authors were not able to show that theta-tACS also enhances the corresponding frequency band. Interestingly, Kirov and colleagues [64] showed that slow oscillatory-tDCS (so-tDCS) applied during encoding is not only able to enhance slow oscillatory activity, but also theta activity. The authors also found an improved encoding performance after so-tDCS compared to sham. They assumed that the improvement is associated with the widespread increase of theta activity promoted by so-tDCS which indicates a considerable role of the respective brain state in encoding, notably theta frequency band. Alongside with findings that stimulation phase-locked on up-states of each SO cycle results in higher memory performance [65], this indicates that theta- stimulation is a promising tool to investigate the functional role of theta in encoding, nocturnal consolidation processes and later retrieval of declarative information. Based on studies of Marshall and colleagues [66,67], I surveyed the impact of so-tDCS on retention performance after a regular nocturnal sleep in different groups of age. In addition to the standard protocol containing a paired associative learning task (PAL; declarative memory), we implemented a visual-spatial task (also a declarative memory task) in order to examine the impact of so-tDCS on a second memory subsystem. Visual stimuli are memorized better than verbal items [68,69], known to decline with ageing [70]. Beside different results for overnight memory consolidation performance in the visual-spatial task (improvement in NAP vs. decline in nocturnal sleep) and no change in performance for PAL in older healthy adults, we found no change in healthy younger subjects for any declarative memory task. PY - 2022 PB - Open Science Framework ER -