Electrocorticography, or ECoG, is a technique that is used to record the electrical activity of the brain via electrodes placed inside the skull. This electrical activity repeatedly rises and falls, and can therefore be represented as a series of waves. All waves have three basic properties: amplitude, frequency and phase. Amplitude describes the height of a wave's peaks (and the depth of its troughs), and frequency defines how many waves are produced per second. The phase of a wave changes from 0° to 360° between two consecutive peaks of that wave and then repeats, similar to the phases of the moon.

Previous studies have shown that brain activity at different frequencies can interact. For instance, neural firing (when nerve impulses are sent from one neuron to the next) is related to ‘high frequency activity’; and the amplitude of high frequency activity can be altered by the phase of other, lower frequency brain activity. It has been suggested that this phenomenon, called ‘phase-amplitude coupling’, might be one way that the brain uses to represent information. This ‘phase coding’ hypothesis has been demonstrated in rodents but is largely untested in humans.

Now, Watrous et al. have explored this hypothesis in epilepsy patients who had ECoG electrodes implanted in their brains for a diagnostic procedure before surgery. These electrodes were used to record brain activity while the patients viewed images from four different categories (houses, scenes, tools and faces).

Watrous et al. found that phase-amplitude coupling occurred in over 40% of the recordings of brain activity. The analysis also revealed that the phase of the lower frequency activity at which the high frequency activity occurred was different for each of the four image categories. This provides support for the phase-coding hypothesis in humans. Furthermore, it suggests that not only how much neural firing occurs but also when (or specifically at what phase) it occurs is important for how the brain represents information. Future studies could now build on this analysis to see if phase-amplitude coupling also supports phase coding and neural representations in other thought processes, such as memory and navigation.

Perceptual representations of the environment are critical to an animal's survival and are believed to occur through coactivated neuronal groups known as cell assemblies. Human neuronal firing (Ekstrom et al., 2007; Kraskov et al., 2007; Chan et al., 2011; Rey et al., 2014) and increases in high-frequency activity (HFA) in the gamma range (above 30 Hz; Jacobs and Kahana, 2009; Jacobs et al., 2012; van Gerven et al., 2013) carry information about perceptual and mnemonic representations. Several recent studies have shown that these two signals are positively correlated (Ray et al., 2008; Manning et al., 2009; Whittingstall and Logothetis, 2009; Miller et al., 2014; Rey et al., 2014; Burke et al., 2015) and are each modulated by the phase of low frequency oscillations (LFO) (O'Keefe and Recce, 1993; Bragin et al., 1995; Skaggs et al., 1996; Canolty et al., 2006; Jacobs et al., 2007; Tort et al., 2009; Axmacher et al., 2010; Rutishauser et al., 2010; McGinn and Valiante, 2014). This modulation is detectable as phase-amplitude coupling (PAC) of gamma amplitude to LFO phase (Buzsaki, 2010; Miller et al., 2014; Aru et al., 2015).

Together, these findings have motivated models positing that LFO phase may organize cell assemblies (Kayser et al., 2012; Lisman and Jensen, 2013; Jensen et al., 2014; Watrous et al., 2015), a form of phase coding (O'Keefe and Recce, 1993). Supporting this view, LFO phase can be used to decode behaviorally relevant information (Belitski et al., 2008, 2010; Fell et al., 2008; Schyns et al., 2011; Lopour et al., 2013; Ng et al., 2013) and phase coded neural activity has been demonstrated in rodents (O'Keefe and Recce, 1993; Skaggs et al., 1996) and monkeys (Kayser et al., 2009; Siegel et al., 2009). Although the PAC observed in humans (Canolty et al., 2006; Axmacher et al., 2010) has been thought to reflect phase-coding, this assumption has yet to be validated because prior studies have not investigated the relation between PAC and decoding from LFO phases.

We have recently proposed that the frequency-specific phase of LFO coordinates neural firing to support neural representations (Watrous and Ekstrom, 2014; Watrous et al., 2015). Here, we tested this prediction, a form of the phase-coding hypothesis in humans, by examining the relation between PAC and neural representations for categories. We analyzed intracranial recordings from 167 electrodes in six patients with pharmaco-resistant epilepsy as they viewed pictures of houses, tools, scenes, and faces. First, we identified PAC on individual electrodes by using a recently developed metric which allows for the characterization of PAC across individual HFA events. On electrodes exhibiting PAC, we then assessed the distinctiveness of each category's phase-coded representation during periods with and without pronounced HFA. Our results suggest that during periods with pronounced HFA, categorical representations can be recovered based on the phase of low-frequency oscillations, supporting the idea of phase-coded neural representations in humans.

We analyzed data from a total of 167 intracranially-recorded EEG electrodes from six patients with pharmaco-resistant epilepsy as they viewed pictures of houses, faces, tools, and outdoor scenes (Figure 1A), testing whether these categories may be represented based on HFA activity at different phases of the LFO (Figure 1B). We first sought to identify electrodes exhibiting PAC and used a data-driven method which allows for the identification of predominant modulating and modulated frequencies (Dvorak and Fenton, 2014; see Figure 1—figure supplement 1 for analysis schematic). Figure 2A–D shows the PAC modulation profile of an example electrode from the basal temporal lobe of patient 3. Figure 2A shows the magnitude of the modulatory signal relative to HFA events (time 0) at different frequencies in the HFA band. PAC is evident as red and blue vertical striping, with maximal modulation of activity at 84 Hz (Figure 2B; ‘HFA event’ marked by arrow in Figure 2C) occurring near the trough of the 2.5 Hz oscillation (see also Figure 2D). Notably, PAC was visible in the raw trace (Figure 2C), the modulatory signal showed rhythmicity (Figure 2D), and there was a clear peak in the power spectrum of both the raw signal and the modulatory signal (Figure 2E, see Figure 2—figure supplement 1 for more examples).

Figure 1 with 1 supplement see all

Download asset Open asset

Figure 2 with 3 supplements see all

Download asset Open asset

Next, we investigated the prevalence of PAC and HFA on each electrode. We found robust evidence for PAC, with at least 20% of electrodes in each patient showing significant PAC (n = 72/167 ‘PAC+’ electrodes, see ‘Materials and methods’ for statistical assessment and inclusion criteria). On PAC+ electrodes, HFA was broadly distributed across trials and time points. Calculating the proportion of trials showing a period of significantly increased HFA (95th percentile, see ‘Materials and methods’ for ‘HFA windows’) on each PAC+ electrode and category, we found that HFA occurred throughout the period of stimulus presentation but increased ∼150 ms after stimulus onset (Figure 2F). 66% of trials had at least one HFA window and this prevalence did not vary by category (Figure 2—figure supplement 2; one-way ANOVA, F(3,284) = 0.6, p > 0.61). These findings converge with prior studies demonstrating increased neural firing and HFA during stimulus presentation and demonstrate pronounced PAC in our paradigm (Canolty et al., 2006; Mormann et al., 2008; Axmacher et al., 2010; Cichy et al., 2014; Rey et al., 2014).

We then determined the frequencies and phases at which PAC is maximal on each PAC+ electrode. Slow-modulating (‘F_phase’) frequencies were significantly clustered in the delta band (0.5–4 Hz; Figure 2G) and HFA modulated frequencies (‘F_amp’) were significantly clustered around slow (∼32 Hz) and fast (∼110 Hz) gamma frequencies (Figure 2H, chi-square goodness of fit test across gamma frequencies, p < 0.004, χ²(22) = 43.6, Cohen's d = 0.77). Furthermore, we found that HFA was typically maximal near the trough of the oscillation (i.e., at 180°; Figure 2I; p < 0.05, Rayleigh test; see Figure 2—figure supplement 1 for additional examples and modulation at other phases).

We next tested if PAC occurs for all four categories, which would be necessary if PAC was related to the representation of categorical information. To this end, we tested each category separately for phase clustering of HFA events at the electrode-specific peak modulatory frequency (‘F_MAX’). This analysis revealed significant clustering for all four categories on 87% (63/72) of PAC+ electrodes (Rayleigh test p < 0.00004, Bonferroni-corrected p < 0.01 for PAC+ electrodes and categories). Phase clustering was observed in each patient and did not vary across categories at F_MAX (one-way ANOVA on resultant vector lengths, F(3,284) = 0.14, p > 0.93). In sum, we found evidence for widespread PAC in each patient at several frequencies and phases of the LFO, similar to single neuron and field potential studies in monkeys (Kayser et al., 2009; Siegel et al., 2009) and humans (Canolty et al., 2006; Jacobs et al., 2007; Axmacher et al., 2010; Maris et al., 2011; Jacobs et al., 2012; van der Meij et al., 2012; Voytek et al., 2015).

HFA occurs at different phases for different categories

Testing the phase-coding hypothesis, we asked if high frequency activity occurred during category-specific phases of the modulatory LFO. Figure 3A shows two traces from an example electrode which are color-coded by the instantaneous phase at F_max. HFA windows (boxes color coded by 1 Hz phase) occurred during different modulatory phases depending on stimulus category. On this electrode, phases extracted during HFA windows were clustered for each category to different phases, resulting in category-specific phase-clustering (Figure 3B). Similar findings were observed in other patients (Figure 3C), and appeared distinct from representations using power or phase (Figure 3—figure supplements 1, 2).

Figure 3 with 3 supplements see all

Download asset Open asset

These findings imply that representations might occur by the category-specific phase at which HFA events occur. In order to further quantify this effect, we developed a simple metric, the difference score (‘DS’), which allowed us to identify the distinctiveness of each category's phase distribution during HFA windows. We applied this metric to the subset of 63 PAC+ electrodes showing significant phase-clustered HFA for each category. This was necessary in order to exclude spurious phase differences between categories occurring in the absence of phase clustering. Across all patients, 78% (49/63) of PAC+ electrodes showed a unique phase-clustering profile for one category compared with each other category (e.g., Figure 1B; DS = 3 for at least one category, p < 10⁻⁹, Watson Williams test, Bonferroni corrected across comparisons). This pattern was consistent both within and across patients, with at least 15% of electrodes in each patient showing these effects (Figure 3D).

We next calculated the average phase difference between categories, expecting this measure to increase with increasing DS. Indeed, categories with larger DSs exhibited larger phase differences with other categories such that maximally distinct representations were 35° phase offset from all other categories (Figure 3E).

As described above, PAC was most likely to occur at the oscillatory trough (Figure 2I and Figure 2—figure supplement 1). Nonetheless, on individual electrodes or for individual categories, HFA could occur at different phases. In fact, across electrodes, phase-coding was equally likely to occur at all phases and for all categories; phase-coded categories were not clustered at particular phases at any level of DS (Rayleigh test, all p > 0.19; Figure 3F) and phase-coding was equally likely for each category (χ²(3) = 1.6, p = 0.64). Thus, a large proportion of PAC+ electrodes also show category-specific phase clustering of HFA events to different phases (Video 1), suggesting that PAC is related to phase-coding (Figure 1B, middle).

Video 1

Download asset

Decoding category identity from HFA event phases

To link these findings more directly to neural coding, we used pattern classification to determine if the phase at which HFA events occur is sufficient to recover categorical information (see ‘Materials and methods’). As expected from the analysis using DS, 42 (25% of all) electrodes showed significant decoding accuracy (using LFO phases during HFA windows as features) compared to category label shuffled surrogates and this proportion was significantly higher than would be expected by chance (p < 10⁻¹⁰, binomial test, chance level: 8.3 electrodes, Cohen's d = 0.6). Next, we assessed whether phase-coding of categorical information indeed depended on HFA, as would be expected if PAC supports phase-coding. We compared decoding accuracy during HFA events to decoding accuracy during randomly selected surrogate events. 19 (11%) electrodes showed significantly higher decoding accuracy during HFA events as compared to random event surrogates, and this proportion was significantly higher than would be expected by chance (p < 0.0003, binomial test, chance = 8.3 electrodes, Cohen's d = 0.22). Moreover, 17 (10%) electrodes showed significant enhancements of decoding accuracy during HFA events relative to both label and event shuffled surrogates, with at least two electrodes in each patient showing this pattern. This proportion of electrodes far exceeded that expected by chance (p < 10⁻¹⁰, binomial test, chance = 0.41 electrodes, Cohen's d = 0.46). These findings complement the above results using DS and indicate that the phase at which HFA events occur carries sufficient information to decode image category, suggesting such information may be a relevant component of the neural code.

We performed several control analyses to rule out alternative explanations. First, if slow oscillatory phase relates to category-specific representations, we expect phase-locking across trials to different categories. We observed significant phase locking on many electrodes to specific categories (Figure 3—figure supplement 3, Rayleigh test, p < 0.000001), similar to previous studies which have identified phase-locked activity (e.g., Fell et al., 2008). Second, we excluded the possibility that our PAC+ or phase-clustering inclusion criteria biased our findings by computing a composite measure of phase representation (PR) on each electrode (see ‘Supplement results’). This analysis again revealed that phase coding is largest on PAC+ electrodes and is enhanced during HFA windows. Third, for comparison with prior PAC methods, we recomputed PAC using the modulation index (MI, Tort et al., 2009) in different low-frequency bands, again finding PAC that was most prevalent in the delta band (Figure 2—figure supplement 3).

Lastly, several models predict that neural processes forming representations will show frequency-specificity (Siegel et al., 2012; Watrous and Ekstrom, 2014; Womelsdorf et al., 2014). We therefore recalculated phase-clustering and DS at the minimum modulatory frequency (F_MIN; see ‘Materials and methods’ and Figure 4—figure supplement 1 for individual subject values) using the same criteria detailed above. As one would expect, on PAC+ electrodes, phase clustering was larger at F_MAX compared to at F_MIN, both on individual electrodes (Figure 4A, B) and at the group level (Figure 4C; paired t-test on resultant vector lengths, t(287) = 8, p < 10⁻¹⁰, Cohen's d = 0.32). Moreover, only 20% (15/72) of PAC+ electrodes showed significant phase-clustering at F_MIN for all 4 categories and only 1 electrode showed category-selective phase-clustering of HFA events. Given that the phase of slower frequencies varies less over time and that we primarily identified F_max at slow frequencies, this result might be biased towards finding enhanced phase clustering at F_max. We therefore recalculated phase clustering across the full range of frequencies (1–12 Hz, 0.1 Hz steps). Again, we found enhanced phase-clustering around 0.5 and 1 Hz (Figure 4—figure supplement 2), but not at adjacent frequencies as would be expected from this alternative account. Taken together, these results support the conclusion that HFA at distinct phases and frequencies reflect representations for different categories.

Figure 4 with 2 supplements see all

Download asset Open asset

We tested the hypothesis that PAC reflects a phase-coding mechanism, measuring both PAC and categorical PR in intracranial recordings from six patients who viewed pictures from different categories. Our analyses show that on a large subset of electrodes showing PAC, the frequency-specific phase at which HFA occurs varies with categorical information. Therefore, to the extent that HFA reflects increases in local neuronal activity (Crone et al., 1998; Miller et al., 2014), our results suggest that neural representations for categories might occur by the phase at which neurons fire. These findings thus provide a novel link between PAC and phase-coded neural representations in humans.

Critically, although PAC and phase-coded representations share some attributes, such as phase-clustering of activity, they are not necessarily identical processes. High frequency activity could occur at particular phases of LFOs, as reflected by PAC, but these phases may not vary with stimulus category (Figure 1B, left). In other words, there could be PAC without phase coding. This is in fact the null hypothesis we have tested and would manifest as PAC with DSs of zero. On the other hand, categorical information may be represented by specific low-frequency phases independent of HFA, leading to DS without phase-clustering across HFA events (PAC; Figure 1B, right). We did not find a complete overlap between PAC+ and phase-coding electrodes, indicating that each can occur in isolation, but instead found a compromise between these extremes (Figure 1B, middle). These results suggest that PAC in many cases reflects phase coding because of the significant overlap between the two phenomena (Video 1).

Phase-coding, in the form of phase-modulated neuronal firing, has been identified in rodents, monkeys, and humans (O'Keefe and Recce, 1993; Skaggs et al., 1996; Jacobs et al., 2007; Kayser et al., 2009; Rutishauser et al., 2010). Although the mechanisms which guide such a neuronal phase preference remain poorly understood, previous studies have found enhanced PAC during learning and memory tasks (Tort et al., 2008; Axmacher et al., 2010; Kendrick et al., 2011; Friese et al., 2013; Lega et al., 2014). Our findings provide a potentially unifying account of these observations, suggesting that PAC may be promoting the formation of phase-coded neural assemblies (Canolty and Knight, 2010; Watrous and Ekstrom, 2014; Watrous et al., 2015). Follow-up studies will need to test this account of PAC as it relates to other putative roles for PAC (Canolty and Knight, 2010; Voytek et al., 2015).

While epilepsy is marked by increased synchronized neuronal activity which could potentially manifest as HFA or PAC, we believe several factors weigh against this interpretation. First, we only analyzed electrodes overlying putatively healthy tissue, typically from the hemisphere contralateral to the epileptic focus, as assessed by our clinical team. Electrodes showing epileptic spiking were systematically removed from our analysis and all analyzed trials were visually inspected for artifacts related to epilepsy. Next, the PAC metric allows for assessment of the modulatory signal. Visual inspection of these signals did not reveal a similarity to epileptic spikes (Figure 2—figure supplement 1). Finally, it seems unlikely that epileptic activity at different phases would systematically differ by category. Similar reasoning excludes saccade-related artifacts (Yuval-Greenberg et al., 2008; Kovatch et al., 2011) as a parsimonious account of our results. We therefore conclude that similar findings would translate into healthy human populations.

Another caveat is that our results provide evidence for categorical phase-coding based on a restricted image set. This was necessary in the present study to maximize the chances of identifying category-selective responses while still ensuring that these responses were generalizable across a few exemplars. Follow-up studies should test the generalizability of these findings using more exemplars within a category and using other categories.

PAC has typically been investigated using pre-defined low and high-frequency filters which may optimize statistical power for detecting PAC but do not adequately deal with the time-resolved nature of cognition (Aru et al., 2015). Here, we leveraged a recent method which can identify PAC and subsequently test mechanistically interesting questions related to the modulation of HFA, such as its temporal profile and its dependence on phase, frequency, and behavioral requirements. Notably, this method may conservatively estimate PAC because it is based on transient increases in HFA, which do not necessarily occur in all cases of PAC. Our findings demonstrate that PAC and large HFA events can be identified and subsequently linked to categorically distinct representations. These results thus extend previous research which has decoded neural representations using either low or high frequency activity (Jacobs and Kahana, 2009; Schyns et al., 2011; van Gerven et al., 2013) and may provide new avenues for decoding the human representational system.

Intriguingly, phase-coding of categorical information extended beyond brain areas associated with higher-order vision. Thus, our findings of category-specificity do not appear to exclusively relate to perception but may also involve other more complex, and idiosyncratic, associations to these stimuli. Our findings are nonetheless in line with prior work (Majima et al., 2014; Yaffe et al., 2014; Zhang et al., 2015) which has found spatially-distributed content-specific representations.

We identified frequency-specific PRs in humans, consistent with a growing body of evidence implicating the relevance of frequency-specific oscillatory activity to human cognition (Daitch et al., 2013; Watrous et al., 2013; Fontolan et al., 2014; Freudenburg et al., 2014). These findings are therefore consistent with models implicating frequency-specific oscillations as central to higher-order cognition (Siegel et al., 2012; Watrous and Ekstrom, 2014; Watrous et al., 2015). It has recently been shown that the frequency of LFOs contributes to several neuronal properties such that relatively slower LFOs lead to decreased firing threshold and increased spike timing variability (Cohen, 2014). It is not immediately clear how this relates to our finding that PAC predominantly occurs with modulating frequencies in the delta band, particularly around 1 Hz. It is possible that our findings reflect the activation of assemblies during ‘up’ states which show a similar frequency profile (Destexhe et al., 2007) or that the applied method of identifying peaks in the spectrum biased our findings to find PAC at lower frequencies.

A third possibility, more likely in our view based on our results indicating multiple modulating frequencies per electrode (Figure 2G), is that the timing of our task (1 image per second with a jittered inter-stimulus interval) partially entrained slow oscillations forming an oscillatory hierarchy (Lakatos et al., 2005). Similarly, our results showing PAC at a variety of phases and frequencies (Maris et al., 2011; van der Meij et al., 2012), particularly near 32 Hz, might reflect a form of ‘nested coupling’ (Kopell et al., 2010) distinct from ‘broadband’ high gamma, which has been suggested to reflect population spiking (Manning et al., 2009; Miller et al., 2014). Future research may clarify this issue by comparing single neuron activity and HFA modulation during different perceptual tasks and by investigating their relation to hierarchical cross-frequency coupling.

To summarize, by identifying electrodes exhibiting both PAC and phase-coded neural representations for categories, our results employing direct brain recordings explicitly link phase-coupled neural activity to phase coding in humans.

1. 1. Aru J
   2. Aru J
   3. Priesemann V
   4. Wibral M
   5. Lana L
   6. Pipa G
   7. Singer W
   8. Vicente R

   (2015) Untangling cross-frequency coupling in neuroscience

   Current Opinion in Neurobiology 31:51–61.

   https://doi.org/10.1016/j.conb.2014.08.002

   • Google Scholar
2. 1. Axmacher N
   2. Elger CE
   3. Fell J

   (2008) Ripples in the medial temporal lobe are relevant for human memory consolidation

   Brain 131:1806–1817.

   https://doi.org/10.1093/brain/awn103

   • Google Scholar
3. 1. Axmacher N
   2. Henseler MM
   3. Jensen O
   4. Weinreich I
   5. Elger CE
   6. Fell J

   (2010) Cross-frequency coupling supports multi-item working memory in the human hippocampus

   Proceedings of the National Academy of Sciences of USA 107:3228–3233.

   https://doi.org/10.1073/pnas.0911531107

   • Google Scholar
4. 1. Belitski A
   2. Gretton A
   3. Magri C
   4. Murayama Y
   5. Montemurro MA
   6. Logothetis NK
   7. Panzeri S

   (2008)

   Low-frequency local field potentials and spikes in primary visual cortex convey independent visual information

   Journal of Neuroscience 28:5696–5709.

   • Google Scholar
5. 1. Belitski A
   2. Panzeri S
   3. Magri C
   4. Logothetis N
   5. Kayser C

   (2010) Sensory information in local field potentials and spikes from visual and auditory cortices: time scales and frequency bands

   Journal of Computational Neuroscience 29:533–545.

   https://doi.org/10.1007/s10827-010-0230-y

   • Google Scholar
6. 1. Berens P

   (2009)

   Circstat: a matlab toolbox for circular statistics

   Journal of Statistical Software 31.

   • Google Scholar
7. 1. Bragin A
   2. Jandó G
   3. Nádasdy Z
   4. Hetke J
   5. Wise K
   6. Buzsáki G

   (1995)

   Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat

   Journal of Neuroscience 15:47–60.

   • Google Scholar
8. 1. Burke JF
   2. Ramayya AG
   3. Kahana MJ

   (2015) Human intracranial high-frequency activity during memory processing: neural oscillations or stochastic volatility

   Current Opinion in Neurobiology 31:104–110.

   https://doi.org/10.1016/j.conb.2014.09.003

   • Google Scholar
9. 1. Buzsaki G

   (2010) Neural syntax: cell assemblies, synapsembles, and readers

   Neuron 68:362–385.

   https://doi.org/10.1016/j.neuron.2010.09.023

   • Google Scholar
10. 1. Canolty RT
    2. Edwards E
    3. Dalal SS
    4. Soltani M
    5. Nagarajan SS
    6. Kirsch HE
    7. Berger MS
    8. Barbaro NM
    9. Knight RT

    (2006) High gamma power is phase-locked to theta oscillations in human neocortex

    Science 313:1626–1628.

    https://doi.org/10.1126/science.1128115

    • Google Scholar
11. 1. Canolty RT
    2. Knight RT

    (2010) The functional role of cross-frequency coupling

    Trends in Cognitive Sciences 14:506–515.

    https://doi.org/10.1016/j.tics.2010.09.001

    • Google Scholar
12. 1. Chan A
    2. Baker J
    3. Eskandar E
    4. Schomer D
    5. Ulbert I
    6. Marinkovic K
    7. Cash S
    8. Halgren E

    (2011) First-pass selectivity for semantic categories in human anteroventral temporal lobe

    Journal of Neuroscience 31:18119–18129.

    https://doi.org/10.1523/JNEUROSCI.3122-11.2011

    • Google Scholar
13. 1. Cichy RM
    2. Pantazis D
    3. Oliva A

    (2014) Resolving human object recognition in space and time

    Nature Neuroscience 17:455–462.

    https://doi.org/10.1038/nn.3635

    • Google Scholar
14. 1. Cohen MX

    (2014) Fluctuations in oscillation frequency control spike timing and coordinate neural networks

    Journal of Neuroscience 34:8988–8998.

    https://doi.org/10.1523/JNEUROSCI.0261-14.2014

    • Google Scholar
15. 1. Crone NE
    2. Miglioretti DL
    3. Gordon B
    4. Lesser RP

    (1998) Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Event-related synchronization in the gamma band

    Brain 121:2301–2315.

    https://doi.org/10.1093/brain/121.12.2301

    • Google Scholar
16. 1. Daitch AL
    2. Sharma M
    3. Roland JL
    4. Astafiev SV
    5. Bundy DT
    6. Gaona CM
    7. Snyder AZ
    8. Shulman GL
    9. Leuthardt EC
    10. Corbetta M

    (2013) Frequency-specific mechanism links human brain networks for spatial attention

    Proceedings of the National Academy of Sciences of USA 110:19585–19590.

    https://doi.org/10.1073/pnas.1307947110

    • Google Scholar
17. 1. Delorme A
    2. Makeig S

    (2004) EEGLAB: an open source toolbox for analysis of single- trial EEG dynamics including independent component analysis

    Journal of Neuroscience Methods 134:9–21.

    https://doi.org/10.1016/j.jneumeth.2003.10.009

    • Google Scholar
18. 1. Destexhe A
    2. Hughes SW
    3. Rudolph M
    4. Crunelli V

    (2007) Are corticothalamic ‘up’ states fragments of wakefulness?

    Trends in Neuroscience 30:334–342.

    https://doi.org/10.1016/j.tins.2007.04.006

    • Google Scholar
19. 1. Dvorak D
    2. Fenton AA

    (2014) Toward a proper estimation of phase-amplitude coupling in neural oscillations

    Journal of Neuroscience Methods 225:42–56.

    https://doi.org/10.1016/j.jneumeth.2014.01.002

    • Google Scholar
20. 1. Ekstrom A
    2. Viskontas I
    3. Kahana M
    4. Jacobs J
    5. Upchurch K
    6. Bookheimer S
    7. Fried I

    (2007) Contrasting roles of neural firing rate and local field potentials in human memory

    Hippocampus 17:606–617.

    https://doi.org/10.1002/hipo.20300

    • Google Scholar
21. 1. Fell J
    2. Ludowig E
    3. Rosburg T
    4. Axmacher N
    5. Elger CE

    (2008) Phase-locking within human mediotemporal lobe predicts memory formation

    Neuroimage 43:410–419.

    https://doi.org/10.1016/j.neuroimage.2008.07.021

    • Google Scholar
22. 1. Fontolan L
    2. Morillon B
    3. Liegeouis-Chauvel C
    4. Giraud AL

    (2014) The contribution of frequency-specific activity to hierarchical information processing in the human auditory cortex

    Nature Communications 5:4694.

    https://doi.org/10.1038/ncomms5694

    • Google Scholar
23. 1. Freudenburg ZV
    2. Gaona CM
    3. Sharma M
    4. Bundy DT
    5. Breshears JD
    6. Pless RB
    7. Leuthardt EC

    (2014) Fast-scale network dynamics in human cortex have specific spectral covariance patterns

    Proceedings of the National Academy of Sciences of USA 111:4602–4607.

    https://doi.org/10.1073/pnas.1311716111

    • Google Scholar
24. 1. Friese U
    2. Köster M
    3. Hassler U
    4. Martens U
    5. Trujillo-Barreto N
    6. Gruber T

    (2013) Successful memory encoding is associated with increased cross-frequency coupling between frontal theta and posterior gamma oscillations in human scalp-recorded EEG

    Neuroimage 66:642–647.

    https://doi.org/10.1016/j.neuroimage.2012.11.002

    • Google Scholar
25. 1. Jacobs J
    2. Kahana M

    (2009) Neural representations of individual stimuli in humans revealed by gamma-band electrocorticographic activity

    Journal of Neuroscience 29:10203–10214.

    https://doi.org/10.1523/JNEUROSCI.2187-09.2009

    • Google Scholar
26. 1. Jacobs J
    2. Kahana M
    3. Ekstrom A
    4. Fried I

    (2007) Brain oscillations control timing of single-neuron activity in humans

    Journal of Neuroscience 27:3839–3844.

    https://doi.org/10.1523/JNEUROSCI.4636-06.2007

    • Google Scholar
27. 1. Jacobs J
    2. Lega B
    3. Anderson C

    (2012) Explaining how brain stimulation can evoke memories

    Journal of Cognitive Neuroscience 24:553–563.

    https://doi.org/10.1162/jocn_a_00170

    • Google Scholar
28. 1. Jensen O
    2. Gips B
    3. Bergmann TO
    4. Bonneford M

    (2014) Temporal coding organized by coupled alpha and gamma oscillations prioritize visual processing

    Trends in Neurosciences 37:357–369.

    https://doi.org/10.1016/j.tins.2014.04.001

    • Google Scholar
29. 1. Kayser C
    2. Ince RA
    3. Panzeri S

    (2012) Analysis of slow (theta) oscillations as a potential temporal reference frame for information coding in sensory cortices

    PLOS Computational Biology 8:e1002717.

    https://doi.org/10.1371/journal.pcbi.1002717

    • Google Scholar
30. 1. Kayser C
    2. Montemurro M
    3. Logothetis N
    4. Panzeri S

    (2009) Spike-phase coding boosts and stabilizes information carried by spatial and temporal spike patterns

    Neuron 61:597–608.

    https://doi.org/10.1016/j.neuron.2009.01.008

    • Google Scholar
31. 1. Kendrick KM
    2. Zhan Y
    3. Fischer H
    4. Nicol AU
    5. Zhang X
    6. Feng J

    (2011) Learning alters theta amplitude, theta-gamma coupling and neuronal synchronization in inferotemporal cortex

    BMC Neuroscience 12:55.

    https://doi.org/10.1186/1471–2202-12–55

    • Google Scholar
32. 1. Kopell N
    2. Kramer MA
    3. Malerba P
    4. Whittington MA

    (2010) Are different rhythms good for different functions?

    Frontiers in Human Neuroscience 4:1–9.

    https://doi.org/10.3389/fnhum.2010.00187

    • Google Scholar
33. 1. Kovatch CK
    2. Tsuchiya N
    3. Kawasaki H
    4. Oya H
    5. Howard MA
    6. Adolphs R

    (2011) Manifestation of ocular-muscle EMG contamination in human intracranial recordings

    Neuroimage 54:213–233.

    https://doi.org/10.1016/j.neuroimage.2010.08.002

    • Google Scholar
34. 1. Kraskov A
    2. Quiroga R
    3. Reddy L
    4. Fried I
    5. Koch C

    (2007) Local field potentials and spikes in the human medial temporal lobe are selective to image category

    Journal of Cognitive Neuroscience 19:479–492.

    https://doi.org/10.1162/jocn.2007.19.3.479

    • Google Scholar
35. 1. Lakatos P
    2. Shah AS
    3. Knuth KH
    4. Ulbert I
    5. Karmos G
    6. Schroeder CE

    (2005) An oscillatory hierarchy controlling neuronal excitability and stimulus processing in the auditory cortex

    Journal of Neurophysiology 94:1904–1911.

    https://doi.org/10.1152/jn.00263.2005

    • Google Scholar
36. 1. Lega B
    2. Burke J
    3. Jacobs J
    4. Kahana MJ

    (2014) Slow-theta-to-gamma phase-amplitude coupling in human hippocampus supports the formation of new episodic memories

    Cerebral Cortex.

    https://doi.org/10.1093/cercor/bhu232

    • Google Scholar
37. 1. Lisman JE
    2. Jensen O

    (2013) The theta-gamma neural code

    Neuron 77:1002–1016.

    https://doi.org/10.1016/j.neuron.2013.03.007

    • Google Scholar
38. 1. Lopour B
    2. Tavassoli A
    3. Fried I
    4. Ringach D

    (2013) Coding of information in the phase of local field potentials within human medial temporal lobe

    Neuron 79:594–606.

    https://doi.org/10.1016/j.neuron.2013.06.001

    • Google Scholar
39. 1. Majima K
    2. Matsuo T
    3. Kawasaki K
    4. Kawai K
    5. Saito N
    6. Hasegawa I
    7. Kamitani Y

    (2014) Decoding visual object categories from temporal correlations of ECoG signals

    Neuroimage 90:74–83.

    https://doi.org/10.1016/j.neuroimage.2013.12.020

    • Google Scholar
40. 1. Manning JR
    2. Jacobs J
    3. Fried I
    4. Kahana MJ

    (2009) Broadband shifts in local field potential power spectra are correlated with single-neuron spiking in humans

    Journal of Neuroscience 29:13613–13620.

    https://doi.org/10.1523/JNEUROSCI.2041-09.2009

    • Google Scholar
41. 1. Maris E
    2. van Vugt M
    3. Kahana M

    (2011) Spatially distributed patterns of oscillatory coupling between high-frequency amplitudes and low-frequency phases in human iEEG

    Neuroimage 54:836–850.

    https://doi.org/10.1016/j.neuroimage.2010.09.029

    • Google Scholar
42. 1. McGinn RJ
    2. Valiante TA

    (2014) Phase-amplitude coupling and interlaminar synchrony are correlated in human neocortex

    The Journal of Neuroscience 34:15923–15930.

    https://doi.org/10.1523/JNEUROSCI.2771-14.2014

    • Google Scholar
43. 1. Mormann F
    2. Kornblith S
    3. Quiroga RQ
    4. Kraskov A
    5. Cerf M
    6. Fried I
    7. Koch C

    (2008) Latency and selectivity of single neurons indicate hierarchical processing in the human medial temporal lobe

    The Journal of Neuroscience  28:8865–8872.

    https://doi.org/10.1523/JNEUROSCI.1640-08.2008

    • Google Scholar
44. 1. Miller KJ
    2. Honey CJ
    3. Hermes D
    4. Rao R
    5. denNijs M
    6. Ojemann JG

    (2014) Broadband changes in the cortical surface potential track activation of functionally diverse neuronal populations

    Neuroimage 85:711–720.

    https://doi.org/10.1016/j.neuroimage.2013.08.070

    • Google Scholar
45. 1. Ng B
    2. Logothetis N
    3. Kayser C

    (2013) EEG phase patterns reflect the selectivity of neural firing

    Cerebral Cortex 23:389–398.

    https://doi.org/10.1093/cercor/bhs031

    • Google Scholar
46. 1. O'Keefe J
    2. Recce ML

    (1993) Phase relationship between hippocampal place units and the EEG theta rhythm

    Hippocampus 3:317–330.

    https://doi.org/10.1002/hipo.450030307

    • Google Scholar
47. 1. Ray S
    2. Crone NE
    3. Niebur PJ
    4. Franaszczuk SS
    5. Hsiao SS

    (2008) Neural correlates of high-gamma oscillations (60–200 Hz) in macaque local field potentials and their potential implications in electrocorticography

    Journal of Neuroscience 28:11526–11536.

    https://doi.org/10.1523/JNEUROSCI.2848-08.2008

    • Google Scholar
48. 1. Rey HG
    2. Fried I
    3. Quian Quiroga R

    (2014) Timing of single-neuron and local field potential responses in the human medial temporal lobe

    Current Biology 24:299–304.

    https://doi.org/10.1016/j.cub.2013.12.004

    • Google Scholar
49. 1. Rizzuto DS
    2. Madsen JR
    3. Bromfield EB
    4. Schulze-Bonhage A
    5. Kahana MJ

    (2006) Human neocortical oscillations exhibit theta phase differences between encoding and retrieval

    Neuroimage 31:1352–1358.

    https://doi.org/10.1016/j.neuroimage.2006.01.009

    • Google Scholar
50. 1. Rutishauser U
    2. Ross I
    3. Mamelak A
    4. Schuman E

    (2010) Human memory strength is predicted by theta-frequency phase-locking of single neurons

    Nature 464:903–907.

    https://doi.org/10.1038/nature08860

    • Google Scholar
51. 1. Schyns P
    2. Thut G
    3. Gross J

    (2011) Cracking the code of oscillatory activity

    PLOS Biology 9:e1001064.

    https://doi.org/10.1371/journal.pbio.1001064

    • Google Scholar
52. 1. Siegel M
    2. Donner TH
    3. Engel AK

    (2012) Spectral fingerprints of large-scale neuronal interactions

    Nature Reviews Neuroscience 13:121–134.

    https://doi.org/10.1038/nrn3137

    • Google Scholar
53. 1. Siegel M
    2. Warden M
    3. Miller E

    (2009) Phase-dependent neuronal coding of objects in short-term memory

    Proceedings of the National Academy of Sciences of USA 106:21341–21346.

    https://doi.org/10.1073/pnas.0908193106

    • Google Scholar
54. 1. Skaggs WE
    2. McNaughton BL
    3. Wilson MA
    4. Barnes CA

    (1996) Theta phase precession in hippocampal neuronal populations and the compression of temporal sequences

    Hippocampus 6:149–172.

    https://doi.org/10.1002/(SICI)1098-1063(1996)6:2<149::AID-HIPO6>3.0.CO;2-K

    • Google Scholar
55. 1. Tort AB
    2. Komorowski RW
    3. Manns JR
    4. Kopell NJ
    5. Eichenbaum H

    (2009) Theta-gamma coupling increases during the learning of item-context associations

    Proceedings of the National Academy of Sciences of USA 106:20942–20947.

    https://doi.org/10.1073/pnas.0911331106

    • Google Scholar
56. 1. van der Meij R
    2. Kahana M
    3. Maris E

    (2012) Phase-amplitude coupling in human electrocorticography is spatially distributed and phase diverse

    Journal of Neuroscience 32:111–123.

    https://doi.org/10.1523/JNEUROSCI.4816-11.2012

    • Google Scholar
57. 1. van Gerven M
    2. Maris E
    3. Sperling M
    4. Sharan A
    5. Litt B
    6. Anderson C
    7. Baltuch G
    8. Jacobs J

    (2013) Decoding the memorization of individual stimuli with direct human brain recordings

    Neuroimage 70:223–232.

    https://doi.org/10.1016/j.neuroimage.2012.12.059

    • Google Scholar
58. 1. Voytek B
    2. Kayser AS
    3. Badre D
    4. Fegen D
    5. Chang EF
    6. Crone NE
    7. Parvizi J
    8. Knight RT
    9. D'Esposito M

    (2015) Oscillatory dynamics coordinating human frontal networks in support of goal maintenance

    Nature Neuroscience 18:1318–1324.

    https://doi.org/10.1038/nn.4071

    • Google Scholar
59. 1. Watrous AJ
    2. Ekstrom AD

    (2014) The spectro-contextual encoding and retrieval theory of episodic memory

    Frontiers in Human Neuroscience 8:75.

    https://doi.org/10.3389/fnhum.2014.00075

    • Google Scholar
60. 1. Watrous AJ
    2. Fell J
    3. Ekstrom AD
    4. Axmacher N

    (2015) More than spikes: common oscillatory mechanisms for content specific neural representations during perception and memory

    Current Opinion in Neurobiology 31:33–39.

    https://doi.org/10.1016/j.conb.2014.07.024

    • Google Scholar
61. 1. Watrous AJ
    2. Tandon N
    3. Conner CR
    4. Pieters T
    5. Ekstrom AD

    (2013) Frequency-specific network connectivity increases underlie accurate spatiotemporal memory retrieval

    Nature Neuroscience 16:349–356.

    https://doi.org/10.1038/nn.3315

    • Google Scholar
62. 1. Whittingstall K
    2. Logothetis NK

    (2009) Frequency-band coupling in surface EEG reflects spiking activity in monkey visual cortex

    Neuron 64:281–289.

    https://doi.org/10.1016/j.neuron.2009.08.016

    • Google Scholar
63. 1. Womelsdorf T
    2. Valiante TA
    3. Sahin NT
    4. Miller KJ
    5. Tiesinga P

    (2014) Dynamic circuit motifs underlying rhythmic gain countrol, gating and integration

    Nature Reviews Neuroscience 17:1031–1039.

    https://doi.org/10.1038/nn.3764

    • Google Scholar
64. 1. Yaffe RB
    2. Kerr MS
    3. Damera S
    4. Sarma SV
    5. Inati SK
    6. Zaghloul KA

    (2014) Reinstatement of distributed cortical oscillations occurs with precise spatiotemporal dynamics during successful memory retrieval

    Proceedings of the National Academy of Science of USA 111:18727–18732.

    https://doi.org/10.1073/pnas.1417017112

    • Google Scholar
65. 1. Yuval-Greenberg S
    2. Tomer O
    3. Keren AS
    4. Nelken I
    5. Deouell LY

    (2008) Transient induced gamma-band responses in EEG as a manifestation of miniature saccades

    Neuron 58:429–441.

    https://doi.org/10.1016/j.neuron.2008.03.027

    • Google Scholar
66. 1. Zhang H
    2. Fell J
    3. Staresina BP
    4. Weber B
    5. Elger CE
    6. Axmacher N

    (2015)

    Gamma power reductions accompany stimulus-specific representations of dynamic events

    Current Biology 25:635–640.

    • Google Scholar

Author details

1. Andrew J Watrous

   Department of Epileptology, University of Bonn, Bonn, Germany

   Contribution

   AJW, Conception and design, Analysis and interpretation of data, Drafting or revising the article, Contributed unpublished essential data or reagents

   For correspondence

   andrew.j.watrous@gmail.com

   Competing interests

   The authors declare that no competing interests exist.

2. Lorena Deuker

   1. Department of Epileptology, University of Bonn, Bonn, Germany
   2. Donders Institute for Brain, Cognition and Behaviour, Radboud University, Nijmegen, Netherlands

   Contribution

   LD, Conception and design, Acquisition of data, Analysis and interpretation of data, Contributed unpublished essential data or reagents

   Competing interests

   The authors declare that no competing interests exist.

3. Juergen Fell

   Department of Epileptology, University of Bonn, Bonn, Germany

   Contribution

   JF, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Nikolai Axmacher

   1. German Center for Neurodegenerative Diseases, Bonn, Germany
   2. Department of Neuropsychology, Institute of Cognitive Neuroscience, Faculty of Psychology, Ruhr University Bochum, Bochum, Germany

   Contribution

   NA, Conception and design, Analysis and interpretation of data, Drafting or revising the article, Contributed unpublished essential data or reagents

   For correspondence

   nikolai.axmacher@ruhr-uni-bochum.de

   Competing interests

   The authors declare that no competing interests exist.

• 3,593

  views

• 723

  downloads

• 50

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.