Local Field Potential Signals Research Articles

Novel Composite Gold-Aluminum Electrode with Application to Neural Recording and Stimulation in Ultrahigh Field Magnetic Resonance Imaging Scanners.

Traditional electrodes used for neural recording and stimulation generate large regions of signal void (no functional MRI signal) when used in ultrahigh field (UHF) MRI scanners. This is a significant disadvantage when simultaneous neural recording/stimulation and fMRI signal acquisition is desired, for example in understanding the functional mechanisms of deep brain stimulation (DBS). In this work, a novel gold-aluminum microwire neural electrode is presented which overcomes this disadvantage. The gold-aluminum design greatly reduces the magnetic susceptibility difference between the electrode and brain tissue leading to significantly reduced regions of signal void. Gold-aluminum microwire samples are imaged at ultrahigh field 16.4 Tesla and compared with gold-only and aluminum-only microwire samples. First, B0 field mapping was used to quantify field distortions at 16.4T and compared with analytical computations in an agarose phantom. The gold-aluminum microwire samples generated substantially less field distortion and signal loss in comparison with gold-only and aluminum-only samples at 16.4T using gradient echo imaging and echo planar imaging sequences. Next, the proposed gold-aluminum electrode was used to successfully record local field potential signals from a rat cortex. The newly proposed gold-aluminum microwire electrode exhibits reduced field distortions and signal loss at 16.4T, a finding which translates to MRI scanners of lower magnetic field strengths as well. The design can be easily reproduced for widespread study of DBS using MRI in animal models. Additionally, the use of non-reactive gold and aluminum materials presents an avenue for translation to human implant applications in the future.

Decoding route selection of pigeon during goal-directed behavior: A joint spike-LFP study

How to reach the goal is one of the core problems that animals must solve to complete goal-directed behavior. Studies have proved the important role of hippocampus (Hp) in spatial navigation and shown that hippocampal neural activities can represent the current location and goal location. However, for the different routes linking these two locations, the neural representation mechanism of the route selection in Hp is not clear. Here, we addressed this question using neural recordings of Hp ensembles and decoding analyses in pigeons performing a goal-directed route selection task known to require Hp participation. The hippocampal spike trains and local field potentials (LFPs) of five pigeons performing the task were acquired and analyzed. We found that the neuron firing rates and power spectrum characteristics in Hp could encode the animal’s route selection during goal-directed behavior, suggesting that the representation of route selection was coherent for hippocampal spike and LFP signals. Decoding results further indicated that joint spike-LFP features resulted in a significant improvement in the representation accuracy of the route selection. These findings of this study will help to understand the encoding mechanism of route selection in goal-directed behavior.

Adaptive changes in local field potential oscillation associated with morphine conditioned place preference in mice

Neural adaptation associated with formation of morphine conditioned place preference remained largely unexplored. This study monitored longitudinal changes in neural signaling during pre-conditioning, conditioning and post-conditioning periods of morphine conditioned place preference (CPP) paradigm for investigation of adaptive mechanisms of opiate addiction. Male Swiss albino mice implanted with intracranial electrodes into the nucleus accumbens (NAc), striatum (STr) and hippocampus (HC) were used for recording of local field potentials (LFPs). Animals received a 10-day schedule for associative learning to pair the specific compartment of the chamber with morphine effects. Exploratory behavior and LFP signals were recorded during pre-conditioning (baseline level), conditioning (day 1, 5 and 10) and post-conditioning (day 1, 4 and 7) periods. Repeated measures one-way ANOVA followed by Tukey test revealed significant increases in number of visit and time spent in morphine compartment during post-conditioning days. Frequency analysis of LFP highlighted the increases in alpha activity (12 – 18 Hz) in the NAc from post-conditioning day 1 until day 7. Moreover, significantly increased coherent activities between the pair of NAc-HC were developed within gamma frequency range (35 – 42 Hz) on morphine conditioning day 10 and disappeared during post-conditioning days. Taken together, these findings emphasized NAc LFP signaling and neural connectivities between the NAc and HC associated with morphine CPP. These adaptive changes might underlie the formation of morphine conditioned place preference and behavioral consequences such as craving and relapse.

Levodopa affects spike and local field synchronisation in the pedunculopontine nucleus of a rat model of Parkinson's disease.

The pedunculopontine nucleus (PPN) undergoes significant anatomic and electrophysiological alterations in Parkinson’s disease (PD), severely impacting locomotion. However, the effect of 6-hydroxydopamine (6-OHDA) lesion and levodopa (L-DOPA) therapy on the relationships between spike activities and local field potential (LFP) within the PPN is not well-understood. Synchronisation between the spike activity of individual neurones and LFP of neuronal ensembles is a crucial problem in the pathogenesis of PD. In this study, LFP signals and spikes in the PPN of rats in control, lesioned, and L-DOPA groups were recorded synchronously with a multi-unit electrical signal acquisition system and analysed for their coherence value, spike-field coherence, and phase-lock relationship. The spike-LFP relationship in the PPN was markedly increased in specific frequency bands because of the 6-OHDA lesion but differed depending on the animal locomotion state and neuronal type. L-DOPA had a limited therapeutic effect on the 6-OHDA-induced increase in the coherence value. Our study demonstrates that the PPN spike-LFP relationship is involved in the pathogenesis of PD and is critical for the effects of L-DOPA, providing a basis for the clinical treatment of refractory PD symptoms.

Behavioral validation of novel high resolution attention decoding method from multi-units & local field potentials

The ability to access brain information in real-time is crucial both for a better understanding of cognitive functions and for the development of therapeutic applications based on brain-machine interfaces. Great success has been achieved in the field of neural motor prosthesis. Progress is still needed in the real-time decoding of higher-order cognitive processes such as covert attention. Recently, we showed that we can track the location of the attentional spotlight using classification methods applied to prefrontal multi-unit activity (MUA) in the non-human primates. Importantly, we demonstrated that the decoded (x,y) attentional spotlight parametrically correlates with the behavior of the monkeys thus validating our decoding of attention. We also demonstrate that this spotlight is extremely dynamic. Here, in order to get closer to non-invasive decoding applications, we extend our previous work to local field potential signals (LFP). Specifically, we achieve, for the first time, high decoding accuracy of the (x,y) location of the attentional spotlight from prefrontal LFP signals, to a degree comparable to that achieved from MUA signals, and we show that this LFP content is predictive of behavior. This LFP attention-related information is maximal in the gamma band (30–250 Hz), peaking between 60 to 120 Hz. In addition, we introduce a novel two-step decoding procedure based on the labelling of maximally attention-informative trials during the decoding procedure. This procedure strongly improves the correlation between our real-time MUA and LFP based decoding and behavioral performance, thus further refining the functional relevance of this real-time decoding of the (x,y) locus of attention. This improvement is more marked for LFP signals than for MUA signals. Overall, this study demonstrates that the attentional spotlight can be accessed from LFP frequency content, in real-time, and can be used to drive high-information content cognitive brain-machine interfaces for the development of new therapeutic strategies.

A stack LSTM structure for decoding continuous force from local field potential signal of primary motor cortex (M1)

BackgroundBrain Computer Interfaces (BCIs) translate the activity of the nervous system to a control signal which is interpretable for an external device. Using continuous motor BCIs, the user will be able to control a robotic arm or a disabled limb continuously. In addition to decoding the target position, accurate decoding of force amplitude is essential for designing BCI systems capable of performing fine movements like grasping. In this study, we proposed a stack Long Short-Term Memory (LSTM) neural network which was able to accurately predict the force amplitude applied by three freely moving rats using their Local Field Potential (LFP) signal.ResultsThe performance of the network was compared with the Partial Least Square (PLS) method. The average coefficient of correlation (r) for three rats were 0.67 in PLS and 0.73 in LSTM based network and the coefficient of determination (R^{2}) were 0.45 and 0.54 for PLS and LSTM based network, respectively. The network was able to accurately decode the force values without explicitly using time lags in the input features. Additionally, the proposed method was able to predict zero-force values very accurately due to benefiting from an output nonlinearity.ConclusionThe proposed stack LSTM structure was able to predict applied force from the LFP signal accurately. In addition to higher accuracy, these results were achieved without explicitly using time lags in input features which can lead to more accurate and faster BCI systems.

Single-Trial Decoding of Motion Direction During Visual Attention From Local Field Potential Signals

Brain-Computer Interface (BCI) based on Local Field Potential (LFP) has recently been developed to restore communication or behavioral functions. LFP provides comprehensive information, due to its stability, robustness, and reach frequency content within the cognitive process. It has been demonstrated that spatial attention can be decoded from brain activity in the visual cortical areas. However, whether motion direction can be decoded from the LFP signal in the primate visual cortex remains uninvestigated, as well as how decoding performance may be influenced by spatial attention. In this paper, these issues were examined by recording LFP from the middle temporal area (MT) of macaque, employing machine learning algorithms. The animal was trained to report a brief direction change in a target stimulus which moved in various directions during a visual attention task. It was found that the LFP-gamma power was able to provide significant information to reliably decode motion direction, compared with other frequency bands, on a single-trial basis. Moreover, the results show that spatial attention leads to enhancements in motion direction discrimination performance. The highest decoding performance was achieved in the high-gamma frequencies (60–120Hz) when targets were presented inside the receptive field in opposite directions. Using a feature selection approach, performance was improved by optimally selecting features where the highest level of participation was observed in the gamma-band. Generally, the results suggest that in the MT area, LFP signals exhibit appreciable information about visual features like motion direction, which could thus be utilized as a control signal for cognitive BCI systems.

Deep Pinsker and James-Stein Neural Networks for Decoding Motor Intentions From Limited Data

Non-parametric regression has been shown to be useful in extracting relevant features from Local Field Potential (LFP) signals for decoding motor intentions. Yet, in many instances, brain-computer interfaces (BCIs) rely on simple classification methods, circumventing deep neural networks (DNNs) due to limited training data. This paper leverages the robustness of several important results in non-parametric regression to harness the potentials of deep learning in limited data setups. We consider a solution that combines Pinsker's theorem as well as its adaptively optimal counterpart due to James-Stein for feature extraction from LFPs, followed by a DNN for classifying motor intentions. We apply our approach to the problem of decoding eye movement intentions from LFPs collected in macaque cortex while the animals perform memory-guided visual saccades to one of eight target locations. The results demonstrate that a DNN classifier trained over the Pinsker features outperforms the benchmark method based on linear discriminant analysis (LDA) trained over the same features.

Exact neural mass model for synaptic-based working memory.

A synaptic theory of Working Memory (WM) has been developed in the last decade as a possible alternative to the persistent spiking paradigm. In this context, we have developed a neural mass model able to reproduce exactly the dynamics of heterogeneous spiking neural networks encompassing realistic cellular mechanisms for short-term synaptic plasticity. This population model reproduces the macroscopic dynamics of the network in terms of the firing rate and the mean membrane potential. The latter quantity allows us to gain insight of the Local Field Potential and electroencephalographic signals measured during WM tasks to characterize the brain activity. More specifically synaptic facilitation and depression integrate each other to efficiently mimic WM operations via either synaptic reactivation or persistent activity. Memory access and loading are related to stimulus-locked transient oscillations followed by a steady-state activity in the β-γ band, thus resembling what is observed in the cortex during vibrotactile stimuli in humans and object recognition in monkeys. Memory juggling and competition emerge already by loading only two items. However more items can be stored in WM by considering neural architectures composed of multiple excitatory populations and a common inhibitory pool. Memory capacity depends strongly on the presentation rate of the items and it maximizes for an optimal frequency range. In particular we provide an analytic expression for the maximal memory capacity. Furthermore, the mean membrane potential turns out to be a suitable proxy to measure the memory load, analogously to event driven potentials in experiments on humans. Finally we show that the γ power increases with the number of loaded items, as reported in many experiments, while θ and β power reveal non monotonic behaviours. In particular, β and γ rhythms are crucially sustained by the inhibitory activity, while the θ rhythm is controlled by excitatory synapses.

Decoding the neural dynamics of free choice in humans.

How do we choose a particular action among equally valid alternatives? Nonhuman primate findings have shown that decision-making implicates modulations in unit firing rates and local field potentials (LFPs) across frontal and parietal cortices. Yet the electrophysiological brain mechanisms that underlie free choice in humans remain ill defined. Here, we address this question using rare intracerebral electroencephalography (EEG) recordings in surgical epilepsy patients performing a delayed oculomotor decision task. We find that the temporal dynamics of high-gamma (HG, 60–140 Hz) neural activity in distinct frontal and parietal brain areas robustly discriminate free choice from instructed saccade planning at the level of single trials. Classification analysis was applied to the LFP signals to isolate decision-related activity from sensory and motor planning processes. Compared with instructed saccades, free-choice trials exhibited delayed and longer-lasting HG activity during the delay period. The temporal dynamics of the decision-specific sustained HG activity indexed the unfolding of a deliberation process, rather than memory maintenance. Taken together, these findings provide the first direct electrophysiological evidence in humans for the role of sustained high-frequency neural activation in frontoparietal cortex in mediating the intrinsically driven process of freely choosing among competing behavioral alternatives.

Modeling Spectral Properties in Stationary Processes of Varying Dimensions with Applications to Brain Local Field Potential Signals

In some applications, it is important to compare the stochastic properties of two multivariate time series that have unequal dimensions. A new method is proposed to compare the spread of spectral information in two multivariate stationary processes with different dimensions. To measure discrepancies, a frequency specific spectral ratio (FS-ratio) statistic is proposed and its asymptotic properties are derived. The FS-ratio is blind to the dimension of the stationary process and captures the proportion of spectral power in various frequency bands. Here we develop a technique to automatically identify frequency bands that carry significant spectral power. We apply our method to track changes in the complexity of a 32-channel local field potential (LFP) signal from a rat following an experimentally induced stroke. At every epoch (a distinct time segment from the duration of the experiment), the nonstationary LFP signal is decomposed into stationary and nonstationary latent sources and the complexity is analyzed through these latent stationary sources and their dimensions that can change across epochs. The analysis indicates that spectral information in the Beta frequency band (12–30 Hertz) demonstrated the greatest change in structure and complexity due to the stroke.

Regularized Kalman filter for brain-computer interfaces using local field potential signals

BackgroundBrain-computer interfaces (BCIs) seek to establish a direct connection from brain to computer, to use in applications such as motor prosthesis control, control of a cursor on the monitor, and so on. Hence, the accuracy of movement decoding from brain signals in BCIs is crucial. The Kalman filter (KF) is often used in BCI systems to decode neural activity and estimate kinetic and kinematic parameters. To use the KF, the state transition matrix, the observation matrix and the covariance matrices of the process and measurement noises must be known in advance, however, in many applications these matrices are not known. Typically, to estimate these parameters, the ordinary least squares method and the sample covariance matrix estimator are used. Our purpose is to enhance the decoding performance of the KF in BCI systems by improving the estimation of the mentioned parameters. New MethodHere, we propose the Regularized Kalman Filter (RKF) which implements two fundamental features: 1) Regularizing the regression estimate of the state equation to improve the estimation of the state transition matrix, and 2) Use of shrinkage method to improve the estimation of the unknown measurement noise covariance matrix. We validated the performance of the proposed method using two datasets of local field potentials obtained from motor cortex of a monkey (Estimation of kinematic parameters during hand movement) and three rats (Estimation of the amount of force applied by hand as a kinetic parameter). ResultsThe results demonstrate that the proposed method outperforms the conventional KF, the KF with feature selection, the Partial least squares, and the Ridge regression approaches.

An ultra-low-power neural signal acquisition analog front-end IC

A four-channel, power-efficient, low-noise neural recording analog front-end (AFE) integrated circuit (IC) comprised of a low-noise amplifier (LNA), a programmable gain amplifier (PGA), and buffers is presented. The proposed AC-coupled capacitive-feedback LNA utilizes the inverter-stacking technique for the core operational transconductance amplifier which achieves four-time reduction in noise at minimal power consumption. The proposed PGA provides additional gain with tunable filtering function where the high-pass cut-off and low-pass cut-off frequencies can be controlled to acquire action potential and local field potential signals either simultaneously or separately. The overall AFE IC has a programmable gain range from 45 ​dB to 63 ​dB and achieves integrated input-referred noise of 3.16 μVRMS within the 10 ​kHz bandwidth, leading to a noise efficiency factor of 2.04 and power efficiency factor of 4.16. The AFE IC is implemented using 180 ​nm CMOS process and consumes 2.82 ​μW per channel powered from the 1-V supply voltage.

Neural mass modeling of slow-fast dynamics of seizure initiation and abortion.

Epilepsy is a dynamic and complex neurological disease affecting about 1% of the worldwide population, among which 30% of the patients are drug-resistant. Epilepsy is characterized by recurrent episodes of paroxysmal neural discharges (the so-called seizures), which manifest themselves through a large-amplitude rhythmic activity observed in depth-EEG recordings, in particular in local field potentials (LFPs). The signature characterizing the transition to seizures involves complex oscillatory patterns, which could serve as a marker to prevent seizure initiation by triggering appropriate therapeutic neurostimulation methods. To investigate such protocols, neurophysiological lumped-parameter models at the mesoscopic scale, namely neural mass models, are powerful tools that not only mimic the LFP signals but also give insights on the neural mechanisms related to different stages of seizures. Here, we analyze the multiple time-scale dynamics of a neural mass model and explain the underlying structure of the complex oscillations observed before seizure initiation. We investigate population-specific effects of the stimulation and the dependence of stimulation parameters on synaptic timescales. In particular, we show that intermediate stimulation frequencies (>20 Hz) can abort seizures if the timescale difference is pronounced. Those results have the potential in the design of therapeutic brain stimulation protocols based on the neurophysiological properties of tissue.

A Low-Noise Chopper Amplifier with Offset and Low-Frequency Noise Compensation DC Servo Loop

The low-frequency and low-amplitude characteristics of neural signals poses challenges to neural signals recording. A low noise amplifier (LNA) plays an important role in the recording front-end. A chopper-stabilized analog front-end amplifier (FEA) for neural signal acquisition is presented in this paper. It solves the noise and offset interference caused by the servo loop in the chopper amplifier structure. The proposed FEA employs a switched-capacitor (SC) integrator with offset and low-frequency noise compensation. Moreover, a dc-blocking impedance is placed for ripple-rejection (RR), and a positive feedback loop is employed to increase input impedance. The proposed circuit is design in a 0.18-µm 1.8-V CMOS process. It achieves a bandwidth of up to 9 kHz for local field potential and action potential signals acquisition. The referred-to-input (RTI) noise is 0.72 µVrms in the 1 Hz~200 Hz frequency band and 3.46 µVrms in the 200 Hz~5 kHz frequency band. The noise effect factor is 0.43 (1 Hz~200 Hz) and 2.08 (200 Hz~5 kHz). CMRR higher than 87 dB and PSRR higher than 85 dB are achieved in the entire pass-band. It consumes a power of 3.96 µW/channel and occupies an area of 0.244 mm2/channel.

Analysis of time-frequency characteristics and coherence of local field potentials during working memory task of rats after high-frequency repeated transcranial magnetic stimulation

Repetitive transcranial magnetic stimulation(rTMS) is a painless and non-invasive method for stimulation and modulation in the field of cognitive neuroscience research and clinical neurological regulation. In this paper, adult Wistar rats were divided into the rTMS group and control group randomly. Rats in the rTMS group were stimulated with 5 Hz rTMS for 14 days, while the rats in the control group did not accept any stimulation. Then, the behavior and local field potentials (LFPs) were recorded synchronously when the rats perform a working memory (WM) task with T-maze. Finally, the time-frequency distribution and coherence characteristics of the LFPs signal in the prefrontal cortex (PFC) during working memory task were analyzed. The results showed that the rats in the rTMS group needed less training days to reach the task correction criterion than the control group (P < 0.05). Compared with the control group, the rTMS group has higher energy (P < 0.01) in θ band (4~12 Hz) and γ band (30~80 Hz). The coherence between the channel pairs decreases as the spatial distance of the channel pairs increases, and the rTMS group exhibits a higher coherence than the control group (P < 0.01). It is concluded that 5 Hz rTMS can improve the excitability of rat prefrontal cortical neurons to a certain extent, and has a positive effect on the working memory ability of normal rats. The results of this paper may provide important theoretical support for further research on the mechanism of action of rTMS on WM.

A user-friendly algorithm for adaptive closed-loop phase-locked stimulation

BackgroundClosed-loop phase-locked stimulation experiments are rare due to the unavailability of user-friendly algorithms and devices. Our goal is to provide an algorithm for the detection of oscillatory activity in local field potentials (LFPs) and phase prediction, which is user-friendly and robust to non-stationarities in LFPs of behaving animals. New methodWe propose an algorithm that only requires specification of the frequency range within which oscillatory episodes are tracked. Frequency-specific detection thresholds and filter parameters are adjusted automatically based on the short-time LFP power spectrum. Estimates of instantaneous frequency and instantaneous phase are used for phase extrapolation, taking advantage of Bayesian estimation. We used real LFP signals, recorded from a variety of different species and different brain areas, as well as artificial LFP signals with known properties to assess the detection and prediction performance of our algorithm and three previously published reference algorithms under various conditions. Results and comparison with existing methodsOur algorithm, while significantly more user-friendly than previous approaches, provides a solid detection and prediction performance over a wide range of realistic conditions and, in many cases, has a longer prediction horizon than the reference algorithms. Due to its ability to adjust to changes in the signal, the algorithm is well-prepared to deal with non-stationarities in oscillation frequency, even in the presence of multiple oscillation components. ConclusionsWe have created a universal algorithm for oscillation detection and phase prediction, which performs well and is user-friendly at the same time, making closed-loop phase-locked stimulation experiments easier to accomplish.

A soft and stretchable bilayer electrode array with independent functional layers for the next generation of brain machine interfaces

Objective. Brain-Machine Interfaces (BMIs) hold great promises for advancing neuroprosthetics, robotics, and for providing treatment options for severe neurological diseases. The objective of this work is the development and in vivo evaluation of electrodes for BMIs that meet the needs to record brain activity at sub-millimeter resolution over a large area of the cortex while being soft and electromechanically robust (i.e. stretchable). Approach. Current electrodes require a trade-off between high spatiotemporal resolution and cortical coverage area. To address the needs for simultaneous high resolution and large cortical coverage, the prototype electrode array developed in this study employs a novel bilayer routing of soft and stretchable lead wires from the recording sites on the surface of the brain (electrocorticography, ECoG) to the data acquisition system. Main results. To validate the recording characteristics, the array was implanted in healthy felines for up to 5 months. Neural signals recorded from both layers of the device showed elevated mid-frequency structures typical of local field potential (LFP) signals that were stable in amplitude over implant duration, and also exhibited consistent frequency-dependent modulation after anesthesia induction by Telazol. Significance. The successful development of a soft and stretchable large-area, high resolution micro ECoG electrode array (lahrμECoG) is an important step to meet the neurotechnological needs of advanced BMI applications.

Scalable surrogate deconvolution for identification of partially-observable systems and brain modeling

Objective. For many biophysical systems, direct measurement of all state-variables, in − vivo is not feasible. Thus, a key challenge in biological modeling and signal processing is to reconstruct the activity and structure of interesting biological systems from indirect measurements. These measurements are often generated by approximately linear time-invariant dynamical interactions with the hidden system and may therefore be described as a convolution of hidden state-variables with an unknown kernel. Approach. In the current work, we present an approach termed surrogate deconvolution, to directly identify such coupled systems (i.e. parameterize models). Surrogate deconvolution reframes certain non linear partially-observable identification problems, which are common in neuroscience/biology, as analytical objectives that are compatible with almost any user-chosen optimization procedure. Main results. We show that the proposed technique is highly scalable, low in computational complexity, and performs competitively with the current gold-standard in partially-observable system estimation: the joint Kalman Filters (Unscented and Extended). We show the benefits of surrogate deconvolution for model identification when applied to simulations of the Local Field Potential and blood oxygen level dependent (BOLD) signal. Lastly, we demonstrate the empirical stability of Hemodynamic Response Function (HRF) kernel estimates for Mesoscale Individualized NeuroDynamic (MINDy) models of individual human brains. The recovered HRF parameters demonstrate reliable individual variation as well as a stereotyped spatial distribution, on average. Significance. These results demonstrate that surrogate deconvolution promises to enhance brain-modeling approaches by simultaneously and rapidly fitting large-scale models of brain networks and the physiological processes which generate neuroscientific measurements (e.g. hemodynamics for BOLD fMRI).

Epi-Intra neural probes with glassy carbon microelectrodes help elucidate neural coding and stimulus encoding in 3D volume of tissue

Objective. In this study, we demonstrate practical applications of a novel 3-dimensional neural probe for simultaneous electrophysiological recordings from the surface of the brain as well as deep intra-cortical tissue. We used this 3D probe to investigate signal propagation mechanisms between neuronal cells and their responses to stimuli in a 3D fashion. Approach. This novel probe leverage 2D thin-film microfabrication technique to combine an epi-cortical (surface) and an intra-cortical (depth) microelectrode arrays (Epi-Intra), that unfold into an origami 3D-like probe during brain implantation. The flexible epi-cortical component conforms to the brain surface while the intra-cortical array is reinforced with stiffer durimide polymer layer for ease of tissue penetration. The microelectrodes are made of glassy carbon material that is biocompatible and has low electrochemical impedance that is important for high fidelity neuronal recordings. These recordings were performed on the auditory region of anesthetized European starling songbirds during playback of conspecific songs as auditory stimuli. Main results. The Epi-Intra probe recorded broadband activity including local field potentials (LFPs) signals as well as single-unit activity and multi-unit activity from both surface and deep brain. The majority of recorded cellular activities were stimulus-locked and exhibited low noise. Notably, while LFPs recorded on surface and depth electrodes did not exhibit strong correlation, composite receptive fields (CRFs)—extracted from individual neuron cells through a non-linear model and that are cell-dependent—were correlated. Significance. These findings demonstrate that CRFs extracted from Epi-Intra recordings are excellent candidates for neural coding and for understanding the relationship between sensory neuronal responses and their stimuli (stimulus encoding). Beyond CRFs, this novel neural probe may enable new spatiotemporal 3D volumetric mapping to address, with cellular resolution, how the brain coordinates function.