The hippocampus has been implicated in episodic and autobiographical memory formation in animal models (Devito and Eichenbaum, 2011; Ergorul and Eichenbaum, 2004; Morris et al., 1982; Squire, 1992) and humans (Squire and Zola-Morgan, 1991; Tulving, 2002). Using electrical stimulation to modulate hippocampal activity to restore memory function has recently become a focus of considerable interest. This idea is motivated to an extent by a long history of research on long-term potentiation (LTP), which is accepted as a neural substrate of learning and memory (Bliss and Lomo, 1973; Kandel and Schwartz, 1982) and can be induced by electrical stimulation to hippocampal afferents, such as the perforant path or Schaffer collaterals (Bliss and Lomo, 1973; Cooke and Bliss, 2006; Larson et al., 1986; Nguyen and Kandel, 1997; Staubli and Lynch, 1987).

Deep brain stimulation (DBS) of hippocampal-related targets, such as the fornix (Hamani et al., 2008; Miller et al., 2015) and the entorhinal region (Suthana et al., 2012), have shown promise for modulating human cognition. However, DBS is commonly applied through pairs of large clinical electrodes (e.g., 1.27 mm diameter), which lack target specificity and, hence, may affect wide and heterogeneous neuronal assemblies in nearby brain regions. More spatially-focused stimulation in the medial temporal lobe (MTL) could be achieved using smaller electrodes (for example, 100 μm diameter microelectrodes), allowing higher anatomical specificity and physiological modulation of brain circuits critical for learning and memory. Moreover, targeting afferent projections may be a more specific and efficient strategy to drive downstream neurons (Riva-Posse et al., 2014), as suggested by recent optogenetic studies (Gradinaru et al., 2009; Rajasethupathy et al., 2016).

Microstimulation mimics effects that drive neurons naturally (Fetsch et al., 2014) and has been shown in animal studies to modulate a variety of specific brain functions, depending on the site of stimulation (Hamani et al., 2008; Histed et al., 2009; Logothetis et al., 2010). For example, in nonhuman primates, stimulation using microelectrodes improved face categorization (Afraz et al., 2006), modulated motion perception with great specificity (Fetsch et al., 2014), and increased the learning rate during a reinforcement learning task (Williams and Eskandar, 2006). In humans, microstimulation of primary visual cortex has been shown to induce phosphenes (Schmidt et al., 1996), and microstimulation of the substantia nigra can influence reinforcement learning (Ramayya et al., 2014). Further evidence of the high specificity that this method of stimulation can bring to human brain modulation has been provided by intraoperative investigations (Histed et al., 2013; Logothetis et al., 2010). Stimulation using microwires presents a novel opportunity to improve selective and natural activation of neuronal circuits in the human brain.

Here, we asked whether application of microstimulation targeted to the entorhinal afferents into hippocampus could enhance declarative memory function in humans. We postulated that afferent stimulation would most directly affect the downstream hippocampal fields, and thus memory function as well. Importantly, LTP induction by high-frequency microstimulation of the perforant path in vivo has been shown to cause reorganization of wide hippocampal and cortical networks in rats (Canals et al., 2009). A theta-burst protocol has been shown to be optimal for inducing LTP (Larson et al., 1986). We therefore hypothesized that theta-burst microstimulation, targeted to brain regions containing afferent fibers to the hippocampus, would improve episodic memory performance in humans.

One crucial aspect of episodic memory is pattern separation, the ability to retrieve the specifics of past events without generalizing to similar or partially overlapping events (Yassa and Stark, 2011). During a task in which subjects studied photos of novel people, microstimulation was applied in a theta-burst pattern. We show that theta-burst microstimulation of the right entorhinal area, applied prior to stimulus onset, enhanced memory specificity for these photographs. That is, during a test phase, subjects performed better at accepting photographs from the encoding stage as previously viewed, concomitant with rejecting similar (lure) photographs.

Participants were thirteen neurosurgical patients with pharmacoresistant epilepsy, implanted with intracranial depth electrodes, who performed a person recognition task (Figure 1, Figure 1—source data 1). During the encoding phase, subjects viewed novel portraits of people. On half of the trials, randomly selected for each participant, electrical stimulation was applied during the fixation period prior to stimulus onset (see Microstimulation Protocol section in Materials and methods) (Figure 1A–B). Stimulation was delivered through a 100 μm diameter microwire that had been targeted to the left or right entorhinal area (two subjects received stimulation on each side (in separate sessions); Figure 1—source data 1, Figure 2). In post hoc analysis, all stimulation microwires were determined to have been located in regions known to include afferent inputs to the hippocampus, including entorhinal white matter (angular bundle), entorhinal gray matter, or subiculum (Figure 2, Figure 2—figure supplements 1–2, Figure 1—source data 1) (Yassa et al., 2010; Zeineh et al., 2017). Because the subiculum contains—in fact, is perforated by—fibers of the perforant path, a major output of the entorhinal cortex, it was included in our sample. After the encoding phase, subjects performed a 30 s distractor task and then were presented with a series of images and asked whether each was 'old’ (presented during the encoding phase) or ‘new.’ For each image presented during the encoding phase, two images were included in the test phase; one, the ‘target’ was the exact same image; the other, the ‘lure,’ was a portrait of a different person who looked similar to the person in the target portrait (Figure 1A,C). These images were presented one at a time, in pseudo-random order, during the test phase. Whenever possible, subjects performed the task more than once, using a new set of images each time, such that a total of 40 experimental sessions were conducted (Figure 1—source data 1; Materials and methods).

Figure 1

Download asset Open asset

Figure 1—source data 1

Participant demographics and stimulation location.

Age, gender, handedness, and hemisphere of language dominance of each of the 13 participants (R: right, L: left, A: ambidextrous, B: bilateral, NA: not available (language dominance was not tested)). Antiepileptic drugs refers to medications taken on the day(s) of the experiments. Also shown are the clinically determined seizure-onset zones and stimulated medial temporal lobe (MTL) regions for each subject. Numbers in parentheses indicate the number of individual experimental sessions in which each subject participated. Participants 6 and 13 participated in multiple sessions and received stimulation on each side, but not concurrently. * indicates a stimulated region that fell within the seizure-onset zone for that participant.

https://doi.org/10.7554/eLife.29515.003

Download elife-29515-fig1-data1-v1.docx

Figure 2 with 2 supplements see all

Download asset Open asset

To determine the behavioral effects of microstimulation, we defined several performance metrics. For each image presented during encoding, we noted whether it was recognized as ‘old’ during the test phase, as well as whether its corresponding lure image was correctly identified as ‘new.’ For each session, we computed the target acceptance rate as the proportion of target images that were correctly identified as old, and the lure rejection rate as the proportion of lure images that were correctly identified as new. Independently, these measures are not fully informative; for example, if a subject has a bias toward answering ‘old’, that would inflate the target acceptance rate while decreasing the lure rejection rate. Thus, a comparison of responses to targets and lures is required. On the whole-session level, we can compute the discrimination index (DI), which is the probability of answering ‘old’ on target images minus the probability of answering ‘old’ on lure images. DI works to subtract out the bias toward one answer or another, but does not explicitly compare performance for specific targets and their corresponding lures. Thus, for each image we assigned a behavioral label of ‘remembered’ or ‘missed’. ‘Remembered’ images were those that were accepted as previously viewed, along with rejection of the corresponding lure. This required not only recognizing the target image but also remembering the details sufficiently to distinguish it from the lure image. We computed the remembered rate as the proportion of target images that met these criteria. All other images were categorized as ‘missed.’

To evaluate how microstimulation affected these behavioral metrics, we computed each metric separately for the subset of stimulated trials and the subset of non-stimulated trials within each session and used generalized estimating equations (GEEs) to model the effects of stimulation condition on these measures. GEEs are a class of generalized linear models that were developed for analyzing repeated measures data (in our case, multiple experiments for an individual subject) and, contrary to the conventional repeated measures ANOVA tests, can handle different numbers of observations per subject and do not assume equal correlations between within-subject observations (Gardiner et al., 2009; Gueorguieva and Krystal, 2004; Hubbard et al., 2010; Subramanian and O'Malley, 2010), thus making this a rigorous statistical approach for our study (see Materials and methods for details)

We hypothesized that stimulation would improve memory specificity, but the effects of stimulation might vary with the precise location of the stimulating electrode. In particular, because previous studies in both human and non-human primates have indicated that facial processing may be lateralized within the hippocampus (Fried et al., 1997; Haxby et al., 1996), we predicted that stimulation within the left or right entorhinal area may be differentially effective. Thus, using GEEs, we modeled the difference in proportion of remembered pictures between stimulated and non-stimulated trials as a function of stimulation hemisphere. We found that there was a significant effect of stimulation hemisphere on the degree to which stimulation changed the proportion of remembered portraits (p=6.17 × 10⁻⁴, Wald ${\mathrm{\chi }}^2$ = 11.72). Specifically, stimulation of the right entorhinal area significantly improved performance (Estimated Mean (EM) = 0.12, 95% CI [0.05, 0.18]), while stimulation of the left entorhinal area had no effect (EM = −0.017, 95% CI [−0.060, 0.025]) (Figure 3A, Figure 3—source data 1, Source code 1). We confirmed that these results were robust to the precise statistical method chosen; even a basic t-test on the difference between mean performance on stimulated and nonstimulated trials per participant revealed an effect of stimulation hemisphere (t(13) = −2.98, p=0.011), and post hoc t-tests indicated that right entorhinal stimulation caused significant improvement (mean change = 0.12 ± 0.04, t(8) = 3.53, p=0.008; effect size: Cohen’s d_z = 1.18, CL = 0.88), while left entorhinal stimulation did not (mean change = −0.02 ± 0.03, t(5) = −0.77, p=0.48) (Figure 3B). Further, out of 9 subjects who received microstimulation in the right entorhinal area, eight had a higher remembered rate for stimulated trials compared to nonstimulated trials, which is more than expected by chance (p=0.020, binomial test). Finally, because our subjects were patients with epilepsy, we wanted to ensure that the effect of stimulation was not driven by epileptic activity. To test this, we applied the GEE model described above to the subset of data that included only sessions in which the stimulating electrode was located outside of the seizure onset zone. Our results remained qualitatively the same, with a strong effect of stimulation hemisphere on behavior, and significant improvement induced by right-sided stimulation (p=9.49 × 10⁻³, Wald ${\mathrm{\chi }}^2$ = 6.73, right EM = 0.11; 95% CI [.029. 184]).

Figure 3

Download asset Open asset

Figure 3—source data 1

Session-level performance data.

https://doi.org/10.7554/eLife.29515.008

Download elife-29515-fig3-data1-v1.xlsx

Figure 3—source data 2

Trial-level performance data.

https://doi.org/10.7554/eLife.29515.009

Download elife-29515-fig3-data2-v1.xlsx

Figure 3—source data 3

Trial by trial analysis of behavioral data.

Each trial’s behavioral outcome was categorized based on whether the viewed portrait was remembered during the test phase or not (total number of trials = 1207 from 13 participants). This metric was modeled as a binary logit response (remembered/missed). Stimulation condition (on/off), stimulation hemisphere (left/right), an interaction term between these two and whether the target image was presented first were specified as factors, which together with normalized trial numbers as a covariate, constituted the model effects. Each factor (row) is quantified by its corresponding (unexponentiated) coefficient (B) and the error associated with it (Std. error and 95% confidence interval), together with whether that term was statistically significant in the model (hypothesis test column).

https://doi.org/10.7554/eLife.29515.010

Download elife-29515-fig3-data3-v1.docx

Although our experimental design included selecting at random which trials in each session received stimulation, we wanted to ensure that our results were independent of any potential trial order effects. We thus modeled the behavioral data on a trial by trial basis while accounting for the trial order as well as whether the target image was presented before the lure image during the test phase (1207 trials). Here, the model included main effects of normalized trial number, whether the target was presented first, stimulation condition (trial received stimulation or not), stimulation hemisphere (left or right), and an interaction term between the last two. We found that the interaction term between the stimulation condition and hemisphere had a strong significant effect on behavior (p=2.19 × 10⁻³, Wald ${\chi }^2$= 9.38), while neither of the order variables were a significant predictor (normalized trial number: p=0.19, presentation of target before lure: p=0.053). Consistent with the previous model, stimulation of the right entorhinal area increased the probability of remembering a portrait to a greater degree than stimulation of the left entorhinal area (Figure 3C; Figure 3—source data 2–3, Source code 2).

Because the order of presentation of target and lure trended strongly toward a behavioral effect—memory was generally better when the target was presented before the lure—we wanted to ensure that the interaction between stimulation condition and hemisphere was independent of whether the target or lure had been presented first. We thus divided our dataset into trials in which the target was presented first (618 trials) and trials in which the lure was presented first (589 trials), and modeled behavioral outcome with normalized trial number, stimulation condition, stimulation hemisphere and an interaction for each dataset. Consistent with the full dataset, we found in each case that normalized trial number was not a significant predictor (target first: p=0.33; lure first: p=0.55), while the interaction between stimulation site and stimulation condition was (target first: p=0.018; lure first: p=0.011).

Finally, we tested whether precise electrode targeting had an effect on the degree to which stimulation improved memory. Our electrodes were localized to three distinct regions containing fibers of the perforant path: the angular bundle (entorhinal white matter), gray matter of the entorhinal cortex (where perforant path fibers originate), and the nearby subiculum (which is perforated by the perforant path prior to its arrival in the hippocampus [Zeineh et al., 2017]). We hypothesized that, because stimulation to the perforant path can induce LTP, stimulating in the region where perforant path fibers are most densely concentrated might be the most effective. Specifically, we compared stimulation in the angular bundle, where perforant path fibers are most densely concentrated (Yassa et al., 2010; Zeineh et al., 2017), to stimulation in either the entorhinal or subicular gray matter, where these fibers are more spread out and fewer axons would be recruited by stimulation. Accordingly, we modeled the proportion of the remembered portraits as a function of two factors: stimulation region (whether the stimulating electrode was in the angular bundle or gray matter) and stimulation hemisphere (left versus right). We found significant main effects of each stimulation hemisphere (p=1.49 × 10⁻⁴, Wald ${\chi }^2$=14.39) and stimulation region (p=0.011, Wald ${\chi }^2$=6.53). The interaction between these factors was not significant (p=0.121, Wald ${\chi }^2$=2.41), indicating that the effects of stimulation on the right side and in the angular bundle each contributed to improved behavioral performance. Specifically, stimulation was most effective for improving memory specificity when the electrodes were positioned in the right angular bundle (EM = 0.16, 95% CI [0.09, 0.22]) (Figure 4A, Figure 3—source data 1; Figure 4—source data 1, Source code 3).

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

Tests of Model Effects for session-level behavioral metrics.

Each behavioral metric was modeled as a linear scale response with stimulation hemisphere (left/right), stimulation region (angular bundle/gray matter), and an interaction term between the two (total number of sessions = 40 from 13 participants). The main effects of stimulation hemisphere and stimulation region, but not the interaction, were significant for remembered rate, discrimination index, and target acceptance rate. The interaction term was significant for lure rejection rate. The Wald Chi-Square statistic, degrees of freedom (df), and significance level (p) are presented for each significant effect.

https://doi.org/10.7554/eLife.29515.013

Download elife-29515-fig4-data1-v1.docx

We next applied this expanded model to test the effect of stimulation on the other behavioral metrics defined above. As expected, DI—also a measure of memory specificity—was improved by right-sided stimulation (p=2.79 × 10⁻³, Wald ${\chi }^2$=8.94) and stimulation in the angular bundle (p=2.53 × 10⁻³, Wald ${\chi }^2$=9.12; interaction term: p=0.15, Wald ${\chi }^2$=2.07; right angular bundle EM = 0.15, 95% CI [0.08, 0.23]) (Figure 4B, Figure 3—source data 1; Figure 4—source data 1, Source code 3). A similar pattern was present in target acceptance, with stimulation in each the right side (p=0.034, Wald ${\chi }^2$=4.47) and the angular bundle (p=2.52 × 10⁻⁵, Wald ${\chi }^2$=17.75) significantly increasing target acceptance rates (right angular bundle EM = 0.074, 95% CI [0.0069. 0.1414]) (Figure 4C, Figure 3—source data 1; Figure 4—source data 1, Source code 3). A different pattern emerged when considering lure rejection rates. Here the interaction between stimulation region and hemisphere was significant (p=0.018, Wald ${\chi }^2$=5.62; right angular bundle EM = 0.080, 95% CI [0.025, 0.134]). The difference appears to be driven by stimulation in the right gray matter showing a trend toward impairing lure rejection (Figure 4D, Figure 3—source data 1; Figure 4—source data 1, Source code 3, Source code 4).

When considering lure rejection and target acceptance together, we found that stimulation on the right side increased the probability of correctly accepting previously viewed target images, regardless of the stimulation region. When right-sided stimulation was applied in the angular bundle, this also led to an increase in appropriately rejecting lure images. When applied in gray matter, however, stimulation led to an increase in incorrectly accepting lure images. Thus, it appears that right angular bundle stimulation improved memory specificity by simultaneously increasing the probability of accepting target images and rejecting lure images. Stimulation of the right gray matter, on the other hand—by increasing acceptance of both targets and lures—introduced a bias towards reporting positive memory, rather than increasing memory per se. Conversely, stimulation in the left entorhinal-adjacent gray matter decreased target acceptance rates and showed a moderate trend toward decreasing lure acceptance rates, introducing a bias toward negative memory. Together these results suggest that in entorhinal/subicular gray matter, stimulation in the left and right hemispheres may introduce opposite biases.

Finally, it should be noted that, although we report uncorrected p-values throughout the results section, all reported statistically significant effects survive correction for multiple comparisons (see Materials and methods), further demonstrating the statistical robustness of our results. We also provide additional analysis in the supplement showing the main effects we report in Figures 3 and 4 when data from the subiculum is excluded (Figure 4—figure supplement 1). The overall results are similar, but further studies will be necessary to examine potential differences between microstimulation of subicular and entorhinal cortices.

We found that microstimulation to the entorhinal area, targeted to maximally affect the entorhinal projections into the hippocampus, resulted in enhancement of memory specificity for images of novel people when the stimulating electrode was in the right hemisphere. This suggests a possible lateralization effect for memory of people, consistent with previous results: the right hippocampus shows increased metabolic activity during the learning of faces (Haxby et al., 1996), and single-unit activity in the right hippocampus encodes different facial expressions (Fried et al., 1997). While direct recordings of neural activity have shown evidence for lateralization, imaging studies have indicated that the encoding of pictorial material involves bilateral activation (Kim, 2011), and lesion studies noted a lack of asymmetry for encoding of non-verbal material (Lee et al., 2002). This discrepancy may indicate a difference between encoding of faces and visual encoding in general, as the latter studies were not specific to images of faces or people. On the other hand, our results indicate that that the left hippocampus is not uninvolved during facial encoding, as stimulation of the left entorhinal area had a behavioral effect, even if it was not to increase memory specificity. Previous findings of bilateral activation, therefore, may indicate only that the two hemispheres are both involved, not that they are performing the same function. Thus, these results warrant further studies into potential lateralization effects, both of function and of the effect of stimulation. It is important to note that our study was conducted in people with epilepsy, a condition which may disproportionately affect one side of the brain and could result in lateralized effects, such as impairment of cognitive skills lateralized to the affected hemisphere. Thus conclusions with respect to lateralization of function should be viewed with caution. However, we have repeated our analysis with data restricted to the sessions in which the stimulating electrode was positioned outside the clinically-determined seizure onset zone, and found qualitatively similar results.

Previous studies in animals have noted the importance of the frequency and mode of stimulation on behavioral outcomes. Specifically, theta-burst stimulation has been used to elicit LTP in a wide range of animal models from hippocampal slices (Nguyen and Kandel, 1997; Staubli and Lynch, 1987) to whole animals (Larson et al., 1986). In the current study, we present evidence supporting the efficacy of theta-burst microstimulation in protocols aimed at human memory improvement. Further research is warranted to determine whether theta-burst stimulation is more effective than other protocols, such as continuous stimulation. We also made no effort in this study to synchronize our stimulation protocols to endogenous theta rhythms in the hippocampus; we hypothesize that the memory effects we observed may be enhanced further by developing a closed-loop stimulation system, allowing the theta-burst pattern to be initiated on specific phases of theta.

Beyond the temporal dynamics of stimulation, we found that the spatial targeting of the stimulating electrode was also a critical factor in determining the effects of stimulation on behavior. Controlling this factor is more feasible with microstimulation than macrostimulation, as it is challenging to confine the spatial extent of delivered current when using macroelectrodes, due to their large contact surface area, wide inter-contact distance for bipolar stimulation, and high magnitude of stimulation current (Figure 2—figure supplement 1). In addition, high frequency stimulation delivered via macroelectrodes has been shown to inhibit nearby neuronal somata, while also providing excitation to axonal projections, indicating that small changes in electrode location could lead to substantially different results (Herrington et al., 2016). For example, stimulating in gray matter could disrupt neuronal computations, whereas electrodes that are confined to white matter may enhance excitatory drive to downstream regions (Afraz et al., 2006; Arcot Desai et al., 2014; Fetsch et al., 2014; Histed et al., 2009). Future work examining the physiological effects of microstimulation in downstream regions, such as the hippocampus, would further enlighten the mechanisms of the differential effects when stimulation is applied in gray matter or the angular bundle.

Recently, studies on the effects of DBS—delivered via macroelectrodes in the entorhinal area—on memory performance have resulted in contradicting outcomes (Jacobs et al., 2016; Suthana et al., 2012). Stimulation delivered in these studies likely affected different neuronal assemblies, which—together with other methodological differences between them—may account for the variability in the reported results. Microstimulation is delivered via a single, small electrode, which eliminates the heterogeneity introduced in macrostimulation studies by variable inter-contact distances and implantation trajectories, and may make the final localization of the electrode easier to compare between research groups; this may increase the ease of replicability of studies using microstimulation, even in the face of other methodological differences.

Additional evidence for the importance of stimulation specificity arises from optogenetic studies in animals. In particular, a transgenic mouse model of early Alzheimer’s disease shows that the more focally targeted an intervention is, the more specific of an effect can be elicited. This study demonstrated that direct optogenetic activation of specific hippocampal memory engram cells resulted in highly specific memory retrieval (Roy et al., 2016). Other optogenetic studies highlight the advantage of afferent stimulation to target downstream neuronal assemblies compared to direct stimulation of gray matter (Gradinaru et al., 2009; Rajasethupathy et al., 2016).

While optogenetics is the state-of-the-art method for targeting neural circuits with precise spatial and temporal resolution (Rajasethupathy et al., 2016), it is not likely to be feasible in humans in the near future. The microstimulation approach may, thus, be the method available in humans that most closely serves a similar purpose. Compared to macroelectrodes, microelectrodes have much smaller electrode contacts and affect a more localized region (Arcot Desai et al., 2014). The lower current levels typically delivered through microelectrodes have also been shown to excite, rather than inhibit, nearby neurons (Afraz et al., 2006; Arcot Desai et al., 2014; Fetsch et al., 2014), as well as those downstream of the nearby axonal projections (Histed et al., 2009). Moreover, high frequency microstimulation of the perforant path that induces LTP in hippocampus has been shown to cause widespread reorganization in hippocampal networks, as well as in cortical areas, such as medial frontal cortex and nucleus accumbens (Canals et al., 2009). These observations are compatible with the efficacy of microstimulation of the perforant path to cause overt behavioral changes in memory performance.

It has been suggested that stimulation in the entorhinal white matter, but not the adjacent gray matter, might lead to beneficial effects on memory (Suthana et al., 2012). Indeed, the results of the current study are consistent with this idea, such that applying microstimulation in the right angular bundle had a significant positive effect on memory for portraits, while applying microstimulation in the right entorhinal gray matter or subiculum did not improve memory, despite the presence of hippocampal afferents in these regions. This may be due to the fact that axons of the perforant path—a common locus for induction of LTP—are most densely concentrated in the angular bundle (Yassa et al., 2010; Zeineh et al., 2017) and hence a larger number would be recruited via stimulation. In parallel, stimulation in the gray matter may more directly impact or interfere with neuronal computations, which could drive the biases observed in our results due to gray matter stimulation. Although we hypothesize that LTP induced via the perforant path is likely responsible for the positive effects of right angular bundle stimulation, it must be borne in mind that the entorhinal region includes several additional fiber tracts—including the direct projections to CA1 via the alvear path, and hippocampal efferents—any of which may be affected by microstimulation, and current imaging techniques do not allow us to separate these various tracts.

Further, it is worth noting that the present study does not address the precise mechanisms by which microstimulation wields its behavioral influence. The distinct patterns observed as a result of stimulation in different hemispheres or different regions offer hints of a complex effect. Future studies investigating how neural signals change in response to microstimulation (and how these changes vary with the precise targeting of the stimulating electrode) will be critical for increasing our understanding, not only of the physiological signatures of microstimulation, but also the microcircuit dynamics underlying memory.

This study presents the first evidence that microstimulation has the potential to improve hippocampal-dependent memory in humans. In addition to the interesting implications for our understanding of microcircuit dynamics of memory, this paves the way to developing new avenues that could one day be used for treating patients with chronic memory impairment. Chronic neural implants that use macrostimulation to treat conditions such as Parkinson’s Disease and Epilepsy have become increasingly common (Fisher and Velasco, 2014; Schuepbach et al., 2013). As the use of neural implants moves toward treating cognitive disorders, one advantage of including microstimulation is the precise spatial targeting it affords, allowing for highly-controlled manipulation of neural circuits. Additionally, microstimulation requires lower levels of current to be delivered, thus allowing the implant to consume considerably less power. Future studies will be required to fully understand the stimulation protocols that provide the best cognitive enhancement for a wide variety of tasks. Large-scale datasets that evaluate the efficacy of macro- vs microstimulation, stimulation region and hemisphere, precise timing of stimulation relative to endogenous brain states, and more, will be required to establish the clinical relevance of stimulation for cognitive enhancement.

At present, there is only one relatively large-scale dataset published to date that attempts to report the effects of (macro) stimulation on memory (Jacobs et al., 2016). Even that study includes only five patients with entorhinal stimulation for one task (spatial memory) and seven for the other (verbal memory). Although that study reported an overall negative effect of stimulation on memory, it is important to note the many differences between it and the present study. In addition to the critical differences between micro- and macrostimulation described above, the tasks in the two studies were quite different. Here we investigated memory specificity in a person recognition task, while in the Jacobs et al., 2016 the tasks required spatial navigation or memorization of a list of words, which may be differentially affected by stimulation. We also accounted for possible confounds of the effectiveness of stimulation, such as the precise location of the stimulating electrode and the hemisphere to receive stimulation. Without accounting for those factors, the positive effects of stimulation could be washed out.

In conclusion, our findings suggest that microstimulation, with its anatomical precision, physiologic-level currents, and action via axonal projections, holds promise for modification of memory circuits and thus for the treatment of memory impairments in people suffering from neurological disorders. More generally, this method may provide a tool for highly specific modulation of neuronal activity and human behavior (Young and Deisseroth, 2017).

1. 1. Afraz SR
   2. Kiani R
   3. Esteky H

   (2006) Microstimulation of inferotemporal cortex influences face categorization

   Nature 442:692–695.

   https://doi.org/10.1038/nature04982

   • PubMed
   • Google Scholar
2. Book

   1. Amaral DG
   2. Insausti R

   (1990)

   The hippocampal formation

   In: Paxinos G, editors. The Human Nervous System. San Diego: Academic Press. pp. 711–755.

   • Google Scholar
3. 1. Arcot Desai S
   2. Gutekunst CA
   3. Potter SM
   4. Gross RE

   (2014) Deep brain stimulation macroelectrodes compared to multiple microelectrodes in rat hippocampus

   Frontiers in Neuroengineering 7:16.

   https://doi.org/10.3389/fneng.2014.00016

   • PubMed
   • Google Scholar
4. 1. Bliss TV
   2. Lomo T

   (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path

   The Journal of Physiology 232:331–356.

   https://doi.org/10.1113/jphysiol.1973.sp010273

   • PubMed
   • Google Scholar
5. 1. Canals S
   2. Beyerlein M
   3. Merkle H
   4. Logothetis NK

   (2009) Functional MRI evidence for LTP-induced neural network reorganization

   Current Biology 19:398–403.

   https://doi.org/10.1016/j.cub.2009.01.037

   • PubMed
   • Google Scholar
6. 1. Cooke SF
   2. Bliss TV

   (2006) Plasticity in the human central nervous system

   Brain 129:1659–1673.

   https://doi.org/10.1093/brain/awl082

   • PubMed
   • Google Scholar
7. 1. Devito LM
   2. Eichenbaum H

   (2011) Memory for the order of events in specific sequences: contributions of the hippocampus and medial prefrontal cortex

   Journal of Neuroscience 31:3169–3175.

   https://doi.org/10.1523/JNEUROSCI.4202-10.2011

   • PubMed
   • Google Scholar
8. Book

   1. Duvernoy HM
   2. Bourgouin P

   (1998)

   The Human Hippocampus: Functional Anatomy, Vascularization, and Serial Sections with MRI

   Berlin: Springer.

   • Google Scholar
9. 1. Ekstrom A
   2. Suthana N
   3. Behnke E
   4. Salamon N
   5. Bookheimer S
   6. Fried I

   (2008) High-resolution depth electrode localization and imaging in patients with pharmacologically intractable epilepsy

   Journal of Neurosurgery 108:812–815.

   https://doi.org/10.3171/JNS/2008/108/4/0812

   • PubMed
   • Google Scholar
10. 1. Ergorul C
    2. Eichenbaum H

    (2004) The hippocampus and memory for "what," "where," and "when"

    Learning & Memory 11:397–405.

    https://doi.org/10.1101/lm.73304

    • PubMed
    • Google Scholar
11. 1. Fetsch CR
    2. Kiani R
    3. Newsome WT
    4. Shadlen MN

    (2014) Effects of cortical microstimulation on confidence in a perceptual decision

    Neuron 83:797–804.

    https://doi.org/10.1016/j.neuron.2014.07.011

    • PubMed
    • Google Scholar
12. 1. Fisher RS
    2. Velasco AL

    (2014) Electrical brain stimulation for epilepsy

    Nature Reviews Neurology 10:261–270.

    https://doi.org/10.1038/nrneurol.2014.59

    • PubMed
    • Google Scholar
13. 1. Fried I
    2. Wilson C
    3. Zhang J
    4. Behnke E
    5. Watson V

    (1993)

    Stereotactic Surgery and Radiosurgery

    149–158, Implantation of depth electrodes for EEG recording, Stereotactic Surgery and Radiosurgery, Madison, WI, Medical Physics Publishing.

    • Google Scholar
14. 1. Fried I
    2. MacDonald KA
    3. Wilson CL

    (1997) Single neuron activity in human hippocampus and amygdala during recognition of faces and objects

    Neuron 18:753–765.

    https://doi.org/10.1016/S0896-6273(00)80315-3

    • PubMed
    • Google Scholar
15. 1. Fried I
    2. Wilson CL
    3. Maidment NT
    4. Engel J
    5. Behnke E
    6. Fields TA
    7. MacDonald KA
    8. Morrow JW
    9. Ackerson L

    (1999) Cerebral microdialysis combined with single-neuron and electroencephalographic recording in neurosurgical patients. Technical note

    Journal of Neurosurgery 91:697–705.

    https://doi.org/10.3171/jns.1999.91.4.0697

    • PubMed
    • Google Scholar
16. 1. Gardiner JC
    2. Luo Z
    3. Roman LA

    (2009) Fixed effects, random effects and GEE: what are the differences?

    Statistics in Medicine 28:221–239.

    https://doi.org/10.1002/sim.3478

    • PubMed
    • Google Scholar
17. 1. Glosser G
    2. Deutsch GK
    3. Cole LC
    4. Corwin J
    5. Saykin AJ

    (1998) Differential lateralization of memory discrimination and response bias in temporal lobe epilepsy patients

    Journal of the International Neuropsychological Society 4:502–511.

    https://doi.org/10.1017/S1355617798455097

    • PubMed
    • Google Scholar
18. 1. Gradinaru V
    2. Mogri M
    3. Thompson KR
    4. Henderson JM
    5. Deisseroth K

    (2009) Optical deconstruction of parkinsonian neural circuitry

    Science 324:354–359.

    https://doi.org/10.1126/science.1167093

    • PubMed
    • Google Scholar
19. 1. Gueorguieva R
    2. Krystal JH

    (2004) Move over anova

    Archives of General Psychiatry 61:310–317.

    https://doi.org/10.1001/archpsyc.61.3.310

    • Google Scholar
20. 1. Gumprecht HK
    2. Widenka DC
    3. Lumenta CB

    (1999) BrainLab vectorvision neuronavigation system: technology and clinical experiences in 131 cases

    Neurosurgery 44:97–104.

    https://doi.org/10.1097/00006123-199901000-00056

    • PubMed
    • Google Scholar
21. 1. Hamani C
    2. McAndrews MP
    3. Cohn M
    4. Oh M
    5. Zumsteg D
    6. Shapiro CM
    7. Wennberg RA
    8. Lozano AM

    (2008) Memory enhancement induced by hypothalamic/fornix deep brain stimulation

    Annals of Neurology 63:119–123.

    https://doi.org/10.1002/ana.21295

    • PubMed
    • Google Scholar
22. Software

    1. Hardin JW

    (2005) Generalized Estimating Equations (GEE), version 2

    John Wiley & Sons, Ltd.

    http://onlinelibrary.wiley.com/doi/
23. 1. Haxby JV
    2. Ungerleider LG
    3. Horwitz B
    4. Maisog JM
    5. Rapoport SI
    6. Grady CL

    (1996) Face encoding and recognition in the human brain

    PNAS 93:922–927.

    https://doi.org/10.1073/pnas.93.2.922

    • PubMed
    • Google Scholar
24. 1. Herrington TM
    2. Cheng JJ
    3. Eskandar EN

    (2016) Mechanisms of deep brain stimulation

    Journal of Neurophysiology 115:19–38.

    https://doi.org/10.1152/jn.00281.2015

    • PubMed
    • Google Scholar
25. 1. Histed MH
    2. Bonin V
    3. Reid RC

    (2009) Direct activation of sparse, distributed populations of cortical neurons by electrical microstimulation

    Neuron 63:508–522.

    https://doi.org/10.1016/j.neuron.2009.07.016

    • PubMed
    • Google Scholar
26. 1. Histed MH
    2. Ni AM
    3. Maunsell JH

    (2013) Insights into cortical mechanisms of behavior from microstimulation experiments

    Progress in Neurobiology 103:115–130.

    https://doi.org/10.1016/j.pneurobio.2012.01.006

    • PubMed
    • Google Scholar
27. 1. Hubbard AE
    2. Ahern J
    3. Fleischer NL
    4. Laan MVder
    5. Lippman SA
    6. Jewell N
    7. Bruckner T
    8. Satariano WA

    (2010) To GEE or Not to GEE

    Epidemiology 21:467–474.

    https://doi.org/10.1097/EDE.0b013e3181caeb90

    • Google Scholar
28. 1. Jacobs J
    2. Miller J
    3. Lee SA
    4. Coffey T
    5. Watrous AJ
    6. Sperling MR
    7. Sharan A
    8. Worrell G
    9. Berry B
    10. Lega B
    11. Jobst BC
    12. Davis K
    13. Gross RE
    14. Sheth SA
    15. Ezzyat Y
    16. Das SR
    17. Stein J
    18. Gorniak R
    19. Kahana MJ
    20. Rizzuto DS

    (2016) Direct electrical stimulation of the human entorhinal region and hippocampus impairs memory

    Neuron 92:983–990.

    https://doi.org/10.1016/j.neuron.2016.10.062

    • Google Scholar
29. 1. Jenkinson M
    2. Smith S

    (2001) A global optimisation method for robust affine registration of brain images

    Medical Image Analysis 5:143–156.

    https://doi.org/10.1016/S1361-8415(01)00036-6

    • PubMed
    • Google Scholar
30. 1. Jenkinson M
    2. Bannister P
    3. Brady M
    4. Smith S

    (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images

    NeuroImage 17:825–841.

    https://doi.org/10.1006/nimg.2002.1132

    • PubMed
    • Google Scholar
31. 1. Kandel ER
    2. Schwartz JH

    (1982) Molecular biology of learning: modulation of transmitter release

    Science 218:433–443.

    https://doi.org/10.1126/science.6289442

    • PubMed
    • Google Scholar
32. 1. Kim H

    (2011) Neural activity that predicts subsequent memory and forgetting: a meta-analysis of 74 fMRI studies

    NeuroImage 54:2446–2461.

    https://doi.org/10.1016/j.neuroimage.2010.09.045

    • PubMed
    • Google Scholar
33. 1. Larson J
    2. Wong D
    3. Lynch G

    (1986) Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation

    Brain Research 368:347–350.

    https://doi.org/10.1016/0006-8993(86)90579-2

    • PubMed
    • Google Scholar
34. 1. Lee TM
    2. Yip JT
    3. Jones-Gotman M

    (2002) Memory deficits after resection from left or right anterior temporal lobe in humans: a meta-analytic review

    Epilepsia 43:283–291.

    https://doi.org/10.1046/j.1528-1157.2002.09901.x

    • PubMed
    • Google Scholar
35. 1. Logothetis NK
    2. Augath M
    3. Murayama Y
    4. Rauch A
    5. Sultan F
    6. Goense J
    7. Oeltermann A
    8. Merkle H

    (2010) The effects of electrical microstimulation on cortical signal propagation

    Nature Neuroscience 13:1283–1291.

    https://doi.org/10.1038/nn.2631

    • PubMed
    • Google Scholar
36. 1. Merrill DR
    2. Bikson M
    3. Jefferys JG

    (2005) Electrical stimulation of excitable tissue: design of efficacious and safe protocols

    Journal of Neuroscience Methods 141:171–198.

    https://doi.org/10.1016/j.jneumeth.2004.10.020

    • PubMed
    • Google Scholar
37. 1. Miller JP
    2. Sweet JA
    3. Bailey CM
    4. Munyon CN
    5. Luders HO
    6. Fastenau PS

    (2015) Visual-spatial memory may be enhanced with theta burst deep brain stimulation of the fornix: a preliminary investigation with four cases

    Brain 138:1833–1842.

    https://doi.org/10.1093/brain/awv095

    • PubMed
    • Google Scholar
38. 1. Morris RG
    2. Garrud P
    3. Rawlins JN
    4. O'Keefe J

    (1982) Place navigation impaired in rats with hippocampal lesions

    Nature 297:681–683.

    https://doi.org/10.1038/297681a0

    • PubMed
    • Google Scholar
39. 1. Nguyen PV
    2. Kandel ER

    (1997) Brief theta-burst stimulation induces a transcription-dependent late phase of LTP requiring cAMP in area CA1 of the mouse hippocampus

    Learning & Memory 4:230–243.

    https://doi.org/10.1101/lm.4.2.230

    • PubMed
    • Google Scholar
40. 1. Pluta J
    2. Yushkevich P
    3. Das S
    4. Wolk D

    (2012) In vivo analysis of hippocampal subfield atrophy in mild cognitive impairment via semi-automatic segmentation of T2-weighted MRI

    Journal of Alzheimer's disease : JAD 31:85–99.

    https://doi.org/10.3233/JAD-2012-111931

    • PubMed
    • Google Scholar
41. 1. Połczyńska MM
    2. Benjamin CF
    3. Moseley BD
    4. Walshaw P
    5. Eliashiv D
    6. Vigil C
    7. Jones M
    8. Bookheimer SY

    (2015) Role of the Wada test and functional magnetic resonance imaging in preoperative mapping of language and memory: two atypical cases

    Neurocase 21:707–720.

    https://doi.org/10.1080/13554794.2014.977300

    • PubMed
    • Google Scholar
42. 1. Rajasethupathy P
    2. Ferenczi E
    3. Deisseroth K

    (2016) Targeting neural circuits

    Cell 165:524–534.

    https://doi.org/10.1016/j.cell.2016.03.047

    • PubMed
    • Google Scholar
43. 1. Ramayya AG
    2. Misra A
    3. Baltuch GH
    4. Kahana MJ

    (2014) Microstimulation of the human substantia nigra alters reinforcement learning

    Journal of Neuroscience 34:6887–6895.

    https://doi.org/10.1523/JNEUROSCI.5445-13.2014

    • PubMed
    • Google Scholar
44. 1. Riva-Posse P
    2. Choi KS
    3. Holtzheimer PE
    4. McIntyre CC
    5. Gross RE
    6. Chaturvedi A
    7. Crowell AL
    8. Garlow SJ
    9. Rajendra JK
    10. Mayberg HS

    (2014) Defining critical white matter pathways mediating successful subcallosal cingulate deep brain stimulation for treatment-resistant depression

    Biological Psychiatry 76:963–969.

    https://doi.org/10.1016/j.biopsych.2014.03.029

    • PubMed
    • Google Scholar
45. 1. Rose TL
    2. Robblee LS

    (1990) Electrical stimulation with Pt electrodes. VIII. Electrochemically safe charge injection limits with 0.2 ms pulses

    IEEE Transactions on Biomedical Engineering 37:1118–1120.

    https://doi.org/10.1109/10.61038

    • PubMed
    • Google Scholar
46. 1. Roy DS
    2. Arons A
    3. Mitchell TI
    4. Pignatelli M
    5. Ryan TJ
    6. Tonegawa S

    (2016) Memory retrieval by activating engram cells in mouse models of early alzheimer's disease

    Nature 531:508–512.

    https://doi.org/10.1038/nature17172

    • PubMed
    • Google Scholar
47. 1. Schlaier JR
    2. Warnat J
    3. Dorenbeck U
    4. Proescholdt M
    5. Schebesch KM
    6. Brawanski A

    (2004) Image fusion of MR images and real-time ultrasonography: evaluation of fusion accuracy combining two commercial instruments, a neuronavigation system and a ultrasound system

    Acta Neurochirurgica 146:271–277.

    https://doi.org/10.1007/s00701-003-0155-6

    • PubMed
    • Google Scholar
48. 1. Schmidt EM
    2. Bak MJ
    3. Hambrecht FT
    4. Kufta CV
    5. O'Rourke DK
    6. Vallabhanath P

    (1996) Feasibility of a visual prosthesis for the blind based on intracortical microstimulation of the visual cortex

    Brain 119:507–522.

    https://doi.org/10.1093/brain/119.2.507

    • PubMed
    • Google Scholar
49. 1. Schuepbach WM
    2. Rau J
    3. Knudsen K
    4. Volkmann J
    5. Krack P
    6. Timmermann L
    7. Hälbig TD
    8. Hesekamp H
    9. Navarro SM
    10. Meier N
    11. Falk D
    12. Mehdorn M
    13. Paschen S
    14. Maarouf M
    15. Barbe MT
    16. Fink GR
    17. Kupsch A
    18. Gruber D
    19. Schneider GH
    20. Seigneuret E
    21. Kistner A
    22. Chaynes P
    23. Ory-Magne F
    24. Brefel Courbon C
    25. Vesper J
    26. Schnitzler A
    27. Wojtecki L
    28. Houeto JL
    29. Bataille B
    30. Maltête D
    31. Damier P
    32. Raoul S
    33. Sixel-Doering F
    34. Hellwig D
    35. Gharabaghi A
    36. Krüger R
    37. Pinsker MO
    38. Amtage F
    39. Régis JM
    40. Witjas T
    41. Thobois S
    42. Mertens P
    43. Kloss M
    44. Hartmann A
    45. Oertel WH
    46. Post B
    47. Speelman H
    48. Agid Y
    49. Schade-Brittinger C
    50. Deuschl G
    51. EARLYSTIM Study Group

    (2013) Neurostimulation for Parkinson's disease with early motor complications

    New England Journal of Medicine 368:610–622.

    https://doi.org/10.1056/NEJMoa1205158

    • PubMed
    • Google Scholar
50. 1. Squire LR
    2. Zola-Morgan S

    (1991) The medial temporal lobe memory system

    Science 253:1380–1386.

    https://doi.org/10.1126/science.1896849

    • PubMed
    • Google Scholar
51. 1. Squire LR

    (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans

    Psychological Review 99:195–231.

    https://doi.org/10.1037/0033-295X.99.2.195

    • PubMed
    • Google Scholar
52. 1. Stark CE
    2. Squire LR

    (2001) Simple and associative recognition memory in the hippocampal region

    Learning & Memory 8:190–197.

    https://doi.org/10.1101/lm.40701

    • PubMed
    • Google Scholar
53. 1. Staubli U
    2. Lynch G

    (1987) Stable hippocampal long-term potentiation elicited by 'theta' pattern stimulation

    Brain Research 435:227–234.

    https://doi.org/10.1016/0006-8993(87)91605-2

    • PubMed
    • Google Scholar
54. 1. Subramanian SV
    2. O'Malley AJ

    (2010) Modeling neighborhood effects: the futility of comparing mixed and marginal approaches

    Epidemiology 21:475–478.

    https://doi.org/10.1097/EDE.0b013e3181d74a71

    • PubMed
    • Google Scholar
55. 1. Suthana NA
    2. Ekstrom AD
    3. Moshirvaziri S
    4. Knowlton B
    5. Bookheimer SY

    (2009) Human hippocampal CA1 involvement during allocentric encoding of spatial information

    Journal of Neuroscience 29:10512–10519.

    https://doi.org/10.1523/JNEUROSCI.0621-09.2009

    • PubMed
    • Google Scholar
56. 1. Suthana N
    2. Haneef Z
    3. Stern J
    4. Mukamel R
    5. Behnke E
    6. Knowlton B
    7. Fried I

    (2012) Memory enhancement and deep-brain stimulation of the entorhinal area

    New England Journal of Medicine 366:502–510.

    https://doi.org/10.1056/NEJMoa1107212

    • PubMed
    • Google Scholar
57. 1. Suthana NA
    2. Parikshak NN
    3. Ekstrom AD
    4. Ison MJ
    5. Knowlton BJ
    6. Bookheimer SY
    7. Fried I

    (2015) Specific responses of human hippocampal neurons are associated with better memory

    PNAS 112:10503–10508.

    https://doi.org/10.1073/pnas.1423036112

    • PubMed
    • Google Scholar
58. 1. Tulving E

    (2002) Episodic memory: from mind to brain

    Annual Review of Psychology 53:1–25.

    https://doi.org/10.1146/annurev.psych.53.100901.135114

    • PubMed
    • Google Scholar
59. Book

    1. Versluis A
    2. Uyttenbroek E

    (2002)

    Exactitudes (1st ed)

    nai010 Publishers.

    • Google Scholar
60. 1. Williams ZM
    2. Eskandar EN

    (2006) Selective enhancement of associative learning by microstimulation of the anterior caudate

    Nature Neuroscience 9:562–568.

    https://doi.org/10.1038/nn1662

    • PubMed
    • Google Scholar
61. 1. Xia M
    2. Wang J
    3. He Y

    (2013) BrainNet Viewer: a network visualization tool for human brain connectomics

    PLoS One 8:e68910.

    https://doi.org/10.1371/journal.pone.0068910

    • PubMed
    • Google Scholar
62. 1. Yassa MA
    2. Muftuler LT
    3. Stark CE

    (2010) Ultrahigh-resolution microstructural diffusion tensor imaging reveals perforant path degradation in aged humans in vivo

    PNAS 107:12687–12691.

    https://doi.org/10.1073/pnas.1002113107

    • PubMed
    • Google Scholar
63. 1. Yassa MA
    2. Stark CE

    (2011) Pattern separation in the hippocampus

    Trends in Neurosciences 34:515–525.

    https://doi.org/10.1016/j.tins.2011.06.006

    • PubMed
    • Google Scholar
64. 1. Young NP
    2. Deisseroth K

    (2017) Cognitive neuroscience: In search of lost time

    Nature 542:173–174.

    https://doi.org/10.1038/nature21497

    • PubMed
    • Google Scholar
65. 1. Yushkevich PA
    2. Wang H
    3. Pluta J
    4. Das SR
    5. Craige C
    6. Avants BB
    7. Weiner MW
    8. Mueller S

    (2010) Nearly automatic segmentation of hippocampal subfields in in vivo focal T2-weighted MRI

    NeuroImage 53:1208–1224.

    https://doi.org/10.1016/j.neuroimage.2010.06.040

    • PubMed
    • Google Scholar
66. 1. Zeineh MM
    2. Palomero-Gallagher N
    3. Axer M
    4. Gräßel D
    5. Goubran M
    6. Wree A
    7. Woods R
    8. Amunts K
    9. Zilles K

    (2017) Direct visualization and mapping of the spatial course of fiber tracts at microscopic resolution in the human hippocampus

    Cerebral Cortex 27:bhw010.

    https://doi.org/10.1093/cercor/bhw010

    • PubMed
    • Google Scholar
67. 1. Zhang Y
    2. Brady M
    3. Smith S

    (2001) Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm

    IEEE Transactions on Medical Imaging 20:45–57.

    https://doi.org/10.1109/42.906424

    • PubMed
    • Google Scholar

Author details

1. Ali S Titiz

   Department of Neurosurgery, University of California, Los Angeles, United States

   Contribution

   Formal analysis, Investigation, Writing—original draft, Writing—review and editing

   Contributed equally with

   Michael R H Hill and Emily A Mankin

   Competing interests

   No competing interests declared

2. Michael R H Hill

   1. Department of Neurosurgery, University of California, Los Angeles, United States
   2. California Institute of Technology, Pasadena, United States

   Contribution

   Conceptualization, Software, Funding acquisition, Validation, Investigation, Methodology, Critically reviewed manuscript

   Contributed equally with

   Ali S Titiz and Emily A Mankin

   Competing interests

   No competing interests declared

3. Emily A Mankin

   Department of Neurosurgery, University of California, Los Angeles, United States

   Contribution

   Data curation, Software, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing

   Contributed equally with

   Ali S Titiz and Michael R H Hill

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2163-913X

4. Zahra M Aghajan

   Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, United States

   Contribution

   Data curation, Formal analysis, Validation, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4245-3975

5. Dawn Eliashiv

   Department of Neurology, University of California, Los Angeles, United States

   Contribution

   Investigation, Methodology, Critically reviewed manuscript

   Competing interests

   No competing interests declared

6. Natalia Tchemodanov

   Department of Neurosurgery, University of California, Los Angeles, United States

   Contribution

   Software, Investigation, Critically reviewed manuscript

   Competing interests

   No competing interests declared

7. Uri Maoz

   1. Department of Neurosurgery, University of California, Los Angeles, United States
   2. California Institute of Technology, Pasadena, United States
   3. Department of Psychology, University of California, Los Angeles, United States

   Present address

   Chapman University, Orange, United States

   Contribution

   Software, Formal analysis, Critically reviewed manuscript

   Competing interests

   No competing interests declared

8. John Stern

   Department of Neurology, University of California, Los Angeles, United States

   Contribution

   Investigation, Methodology, Critically reviewed manuscript

   Competing interests

   No competing interests declared

9. Michelle E Tran

   Department of Neurosurgery, University of California, Los Angeles, United States

   Contribution

   Data curation, Formal analysis, Investigation, Project administration, Critically reviewed manuscript

   Competing interests

   No competing interests declared

10. Peter Schuette

    Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, United States

    Contribution

    Formal analysis, Visualization, Critically reviewed manuscript

    Competing interests

    No competing interests declared

11. Eric Behnke

    Department of Neurosurgery, University of California, Los Angeles, United States

    Contribution

    Visualization, Methodology, Critically reviewed manuscript

    Competing interests

    No competing interests declared

12. Nanthia A Suthana

    1. Department of Neurosurgery, University of California, Los Angeles, United States
    2. Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, United States
    3. Department of Psychology, University of California, Los Angeles, United States

    Contribution

    Conceptualization, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Writing—review and editing

    Contributed equally with

    Itzhak Fried

    Competing interests

    No competing interests declared

13. Itzhak Fried

    1. Department of Neurosurgery, University of California, Los Angeles, United States
    2. Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, United States

    Contribution

    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

    Contributed equally with

    Nanthia A Suthana

    For correspondence

    ifried@mednet.ucla.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-5962-2678

• 4,942

  views

• 681

  downloads

• 95

  citations

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