Publications

Export 1812 results:
Sort by:
2016
Lalazar, H, Abbott LF, Vaadia E.  2016.  Tuning Curves for Arm Posture Control in Motor Cortex Are Consistent with Random Connectivity.. PLoS computational biology. 12(5):e1004910. Abstract
Neuronal responses characterized by regular tuning curves are typically assumed to arise from structured synaptic connectivity. However, many responses exhibit both regular and irregular components. To address the relationship between tuning curve properties and underlying circuitry, we analyzed neuronal activity recorded from primary motor cortex (M1) of monkeys performing a 3D arm posture control task and compared the results with a neural network model. Posture control is well suited for examining M1 neuronal tuning because it avoids the dynamic complexity of time-varying movements. As a function of hand position, the neuronal responses have a linear component, as has previously been described, as well as heterogeneous and highly irregular nonlinearities. These nonlinear components involve high spatial frequencies and therefore do not support explicit encoding of movement parameters. Yet both the linear and nonlinear components contribute to the decoding of EMG of major muscles used in the task. Remarkably, despite the presence of a strong linear component, a feedforward neural network model with entirely random connectivity can replicate the data, including both the mean and distributions of the linear and nonlinear components as well as several other features of the neuronal responses. This result shows that smoothness provided by the regularity in the inputs to M1 can impose apparent structure on neural responses, in this case a strong linear (also known as cosine) tuning component, even in the absence of ordered synaptic connectivity.
Lee, J, Joshua M, Medina JF, Lisberger SG.  2016.  Signal, Noise, and Variation in Neural and Sensory-Motor Latency.. Neuron. Abstract
Analysis of the neural code for sensory-motor latency in smooth pursuit eye movements reveals general principles of neural variation and the specific origin of motor latency. The trial-by-trial variation in neural latency in MT comprises a shared component expressed as neuron-neuron latency correlations and an independent component that is local to each neuron. The independent component arises heavily from fluctuations in the underlying probability of spiking, with an unexpectedly small contribution from the stochastic nature of spiking itself. The shared component causes the latency of single-neuron responses in MT to be weakly predictive of the behavioral latency of pursuit. Neural latency deeper in the motor system is more strongly predictive of behavioral latency. A model reproduces both the variance of behavioral latency and the neuron-behavior latency correlations in MT if it includes realistic neural latency variation, neuron-neuron latency correlations in MT, and noisy gain control downstream of MT.
Atlan, G, Terem A, Peretz-Rivlin N, Groysman M, Citri A.  2016.  Mapping synaptic cortico-claustral connectivity in the mouse.. The Journal of comparative neurology. Abstract
The claustrum is an intriguing brain structure, featuring the highest connectivity per regional volume in the brain. It is a thin and elongated structure enclosed between the striatum and the insular cortex, with widespread reciprocal connections with the sensory modalities and prefrontal cortices. Retinotopic and somatotopic organizations have been described in the claustrum, and anatomical studies in cats, monkeys, and rats have demonstrated topographic organization of cortico-claustral connections. In this study, we mapped the projections from cortical modalities (visual, auditory, somatosensory, motor and olfactory), and prefrontal regions (anterior cingulate cortex and orbitofrontal cortex) to the claustrum in mice. Utilizing expression of a virally-encoded synaptic anterograde tracer, AAV-SynaptoTag, followed by 3-dimensional reconstruction of the cortical projections, we performed a comprehensive study of the organization of these projections within the mouse claustrum. Our results clearly demonstrate a dorsoventral laminar organization of projections from the sensory cortices to the claustrum, whereas frontal inputs are more extensive and overlap with the inputs from the sensory cortices. In addition, we find evidence in support of a core/shell organization of the claustrum. We propose that the overlap between the frontal inputs and the inputs from the sensory modalities may underlie executive regulation of the communication between the claustrum and the cortical modalities. This article is protected by copyright. All rights reserved.
Kronman, A, Joskowicz L.  2016.  A geometric method for the detection and correction of segmentation leaks of anatomical structures in volumetric medical images.. International journal of computer assisted radiology and surgery. 11(3):369-80. Abstract
Patient-specific models of anatomical structures and pathologies generated from volumetric medical images play an increasingly central role in many aspects of patient care. A key task in generating these models is the segmentation of anatomical structures and pathologies of interest. Although numerous segmentation methods are available, they often produce erroneous delineations that require time-consuming modifications.
Deouell, LY, Grill-Spector K, Malach R, Murray MM, Rossion B.  2016.  Introduction to the special issue on functional selectivity in perceptual and cognitive systems--a tribute to Shlomo Bentin (1946-2012).. Neuropsychologia. 83:1-4.
Israel, Z, Bergman H.  2016.  Location, location, location: Validating the position of deep brain stimulation electrodes.. Movement disorders : official journal of the Movement Disorder Society. 31(3):259.
Dürschmid, S, Edwards E, Reichert C, Dewar C, Hinrichs H, Heinze H-J, Kirsch HE, Dalal SS, Deouell LY, Knight RT.  2016.  Hierarchy of prediction errors for auditory events in human temporal and frontal cortex.. Proceedings of the National Academy of Sciences of the United States of America. 113(24):6755-60. Abstract
Predictive coding theories posit that neural networks learn statistical regularities in the environment for comparison with actual outcomes, signaling a prediction error (PE) when sensory deviation occurs. PE studies in audition have capitalized on low-frequency event-related potentials (LF-ERPs), such as the mismatch negativity. However, local cortical activity is well-indexed by higher-frequency bands [high-γ band (Hγ): 80-150 Hz]. We compared patterns of human Hγ and LF-ERPs in deviance detection using electrocorticographic recordings from subdural electrodes over frontal and temporal cortices. Patients listened to trains of task-irrelevant tones in two conditions differing in the predictability of a deviation from repetitive background stimuli (fully predictable vs. unpredictable deviants). We found deviance-related responses in both frequency bands over lateral temporal and inferior frontal cortex, with an earlier latency for Hγ than for LF-ERPs. Critically, frontal Hγ activity but not LF-ERPs discriminated between fully predictable and unpredictable changes, with frontal cortex sensitive to unpredictable events. The results highlight the role of frontal cortex and Hγ activity in deviance detection and PE generation.
Breska, A, Deouell LY.  2016.  When Synchronizing to Rhythms Is Not a Good Thing: Modulations of Preparatory and Post-Target Neural Activity When Shifting Attention Away from On-Beat Times of a Distracting Rhythm.. The Journal of neuroscience : the official journal of the Society for Neuroscience. 36(27):7154-66. Abstract
Environmental rhythms potently drive predictive resource allocation in time, typically leading to perceptual and motor benefits for on-beat, relative to off-beat, times, even if the rhythmic stream is not intentionally used. In two human EEG experiments, we investigated the behavioral and electrophysiological expressions of using rhythms to direct resources away from on-beat times. This allowed us to distinguish goal-directed attention from the automatic capture of attention by rhythms. The following three conditions were compared: (1) a rhythmic stream with targets appearing frequently at a fixed off-beat position; (2) a rhythmic stream with targets appearing frequently at on-beat times; and (3) a nonrhythmic stream with matched target intervals. Shifting resources away from on-beat times was expressed in the slowing of responses to on-beat targets, but not in the facilitation of off-beat targets. The shifting of resources was accompanied by anticipatory adjustment of the contingent negative variation (CNV) buildup toward the expected off-beat time. In the second experiment, off-beat times were jittered, resulting in a similar CNV adjustment and also in preparatory amplitude reduction of beta-band activity. Thus, the CNV and beta activity track the relevance of time points and not the rhythm, given sufficient incentive. Furthermore, the effects of task relevance (appearing in a task-relevant vs irrelevant time) and rhythm (appearing on beat vs off beat) had additive behavioral effects and also dissociable neural manifestations in target-evoked activity: rhythm affected the target response as early as the P1 component, while relevance affected only the later N2 and P3. Thus, these two factors operate by distinct mechanisms.
Wigderson, E, Nelken I, Yarom Y.  2016.  Early multisensory integration of self and source motion in the auditory system.. Proceedings of the National Academy of Sciences of the United States of America. 113(29):8308-13. Abstract
Discriminating external from self-produced sensory inputs is a major challenge for brains. In the auditory system, sound localization must account for movements of the head and ears, a computation likely to involve multimodal integration. Principal neurons (PNs) of the dorsal cochlear nucleus (DCN) are known to be spatially selective and to receive multimodal sensory information. We studied the responses of PNs to body rotation with or without sound stimulation, as well as to sound source rotation with stationary body. We demonstrated that PNs are sensitive to head direction, and, in the presence of sound, they differentiate between body and sound source movement. Thus, the output of the DCN provides the brain with enough information to disambiguate the movement of a sound source from an acoustically identical relative movement produced by motion of the animal.
Wigderson, E, Nelken I, Yarom Y.  2016.  Early multisensory integration of self and source motion in the auditory system.. Proc Natl Acad Sci U S A. 113(29): 8308–8313. Abstract
Discriminating external from self-produced sensory inputs is a major challenge for brains. In the auditory system, sound localization must account for movements of the head and ears, a computation likely to involve multimodal integration. Principal neurons (PNs) of the dorsal cochlear nucleus (DCN) are known to be spatially selective and to receive multimodal sensory information. We studied the responses of PNs to body rotation with or without sound stimulation, as well as to sound source rotation with stationary body. We demonstrated that PNs are sensitive to head direction, and, in the presence of sound, they differentiate between body and sound source movement. Thus, the output of the DCN provides the brain with enough information to disambiguate the movement of a sound source from an acoustically identical relative movement produced by motion of the animal.
Jaffe-Dax, S, Lieder I, Biron T, Ahissar M.  2016.  Dyslexics' usage of visual priors is impaired.. Journal of vision. 16(9):10. Abstract
Human perception benefits substantially from familiarity, via the formation of effective predictions of the environment's pattern of stimulation. Basic stimulation characteristics are automatically retrieved and integrated into our perception. A quantitatively measurable manifestation of the integration of priors is known as "contraction to the mean"; i.e., perception is biased toward the experienced mean. We previously showed that in the context of auditory discrimination, the magnitude of this bias is smaller among dyslexic individuals than among good readers matched for age and general reasoning skills. Here we examined whether a similarly reduced contraction characterizes dyslexics' behavior on serial visual tasks. Using serial spatial frequency discrimination tasks, we found that dyslexics' bias toward the experiment's mean spatial frequency was smaller than that observed for the controls. Thus, dyslexics' difficulties in automatic detection and integration of stimulus statistics are domain-general. These difficulties are likely to impede the acquisition of reading expertise.
Peer, M, Abboud S, Hertz U, Amedi A, Arzy S.  2016.  Intensity-based masking: A tool to improve functional connectivity results of resting-state fMRI.. Human brain mapping. 37(7):2407-18. Abstract
Seed-based functional connectivity (FC) of resting-state functional MRI data is a widely used methodology, enabling the identification of functional brain networks in health and disease. Based on signal correlations across the brain, FC measures are highly sensitive to noise. A somewhat neglected source of noise is the fMRI signal attenuation found in cortical regions in close vicinity to sinuses and air cavities, mainly in the orbitofrontal, anterior frontal and inferior temporal cortices. BOLD signal recorded at these regions suffers from dropout due to susceptibility artifacts, resulting in an attenuated signal with reduced signal-to-noise ratio in as many as 10% of cortical voxels. Nevertheless, signal attenuation is largely overlooked during FC analysis. Here we first demonstrate that signal attenuation can significantly influence FC measures by introducing false functional correlations and diminishing existing correlations between brain regions. We then propose a method for the detection and removal of the attenuated signal ("intensity-based masking") by fitting a Gaussian-based model to the signal intensity distribution and calculating an intensity threshold tailored per subject. Finally, we apply our method on real-world data, showing that it diminishes false correlations caused by signal dropout, and significantly improves the ability to detect functional networks in single subjects. Furthermore, we show that our method increases inter-subject similarity in FC, enabling reliable distinction of different functional networks. We propose to include the intensity-based masking method as a common practice in the pre-processing of seed-based functional connectivity analysis, and provide software tools for the computation of intensity-based masks on fMRI data. Hum Brain Mapp 37:2407-2418, 2016. © 2016 Wiley Periodicals, Inc.
Deffains, M, Iskhakova L, Bergman H.  2016.  Stop and Think about Basal Ganglia Functional Organization: The Pallido-Striatal "Stop" Route.. Neuron. 89(2):237-9. Abstract
The "arkypallidal" neurons of the globus pallidus (external segment) emit feedback GABAergic projections to the striatum. In this issue of Neuron, Mallet et al. (2016) show that "arkypallidal" neurons provide a Stop signal, suppressing the development of Go-related striatal activity.
Marciano-Romm, D, Romm A, Bourgeois-Gironde S, Deouell LY.  2016.  The Alternative Omen Effect: Illusory negative correlation between the outcomes of choice options.. Cognition. 146:324-38. Abstract
In situations of choice between uncertain options, one might get feedback on both the outcome of the chosen option and the outcome of the unchosen option ("the alternative"). Extensive research has shown that when both outcomes are eventually revealed, the alternative's outcome influences the way people evaluate their own outcome. In a series of experiments, we examined whether the outcome of the alternative plays an additional role in the decision-making process by creating expectations regarding the outcome of the chosen option. Specifically, we hypothesized that people see a good (bad) alternative's outcome as a bad (good) sign regarding their own outcome when the two outcomes are in fact uncorrelated, a phenomenon we call the "Alternative Omen Effect" (ALOE). Subjects had to repeatedly choose between two boxes, the outcomes of which were then sequentially revealed. In Experiments 1 and 2 the alternative's outcome was presented first, and we assessed the individual's prediction of their own outcome. In Experiment 3, subjects had to predict the alternative's outcome after seeing their own. We find that even though the two outcomes were in fact uncorrelated, people tended to see a good (bad) alternative outcome as a bad (good) sign regarding their own outcome. Importantly, this illusory negative correlation affected subsequent behavior and led to irrational choices. Furthermore, the order of presentation was critical: when the outcome of the chosen option was presented first, the effect disappeared, suggesting that this illusory negative correlation is influenced by self-relevance. We discuss the possible sources of this illusory correlation as well as its implications for research on counterfactual thinking.
Tal, Z, Geva R, Amedi A.  2016.  The origins of metamodality in visual object area LO: Bodily topographical biases and increased functional connectivity to S1.. NeuroImage. 127:363-75. Abstract
Recent evidence from blind participants suggests that visual areas are task-oriented and sensory modality input independent rather than sensory-specific to vision. Specifically, visual areas are thought to retain their functional selectivity when using non-visual inputs (touch or sound) even without having any visual experience. However, this theory is still controversial since it is not clear whether this also characterizes the sighted brain, and whether the reported results in the sighted reflect basic fundamental a-modal processes or are an epiphenomenon to a large extent. In the current study, we addressed these questions using a series of fMRI experiments aimed to explore visual cortex responses to passive touch on various body parts and the coupling between the parietal and visual cortices as manifested by functional connectivity. We show that passive touch robustly activated the object selective parts of the lateral-occipital (LO) cortex while deactivating almost all other occipital-retinotopic-areas. Furthermore, passive touch responses in the visual cortex were specific to hand and upper trunk stimulations. Psychophysiological interaction (PPI) analysis suggests that LO is functionally connected to the hand area in the primary somatosensory homunculus (S1), during hand and shoulder stimulations but not to any of the other body parts. We suggest that LO is a fundamental hub that serves as a node between visual-object selective areas and S1 hand representation, probably due to the critical evolutionary role of touch in object recognition and manipulation. These results might also point to a more general principle suggesting that recruitment or deactivation of the visual cortex by other sensory input depends on the ecological relevance of the information conveyed by this input to the task/computations carried out by each area or network. This is likely to rely on the unique and differential pattern of connectivity for each visual area with the rest of the brain.
Deffains, M, Iskhakova L, Katabi S, Haber SN, Israel Z, Bergman H.  2016.  Subthalamic, not striatal, activity correlates with basal ganglia downstream activity in normal and parkinsonian monkeys.. eLife. 5 Abstract
The striatum and the subthalamic nucleus (STN) constitute the input stage of the basal ganglia (BG) network and together innervate BG downstream structures using GABA and glutamate, respectively. Comparison of the neuronal activity in BG input and downstream structures reveals that subthalamic, not striatal, activity fluctuations correlate with modulations in the increase/decrease discharge balance of BG downstream neurons during temporal discounting classical condition task. After induction of parkinsonism with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), abnormal low beta (8-15 Hz) spiking and local field potential (LFP) oscillations resonate across the BG network. Nevertheless, LFP beta oscillations entrain spiking activity of STN, striatal cholinergic interneurons and BG downstream structures, but do not entrain spiking activity of striatal projection neurons. Our results highlight the pivotal role of STN divergent projections in BG physiology and pathophysiology and may explain why STN is such an effective site for invasive treatment of advanced Parkinson's disease and other BG-related disorders.
Sabbah, N, Authié CN, Sanda N, Mohand-Saïd S, Sahel J-A, Safran AB, Habas C, Amedi A.  2016.  Increased functional connectivity between language and visually deprived areas in late and partial blindness.. NeuroImage. 136:162-73. Abstract
In the congenitally blind, language processing involves visual areas. In the case of normal visual development however, it remains unclear whether later visual loss induces interactions between the language and visual areas. This study compared the resting-state functional connectivity (FC) of retinotopic and language areas in two unique groups of late visually deprived subjects: (1) blind individuals suffering from retinitis pigmentosa (RP), (2) RP subjects without a visual periphery but with preserved central "tunnel vision", both of whom were contrasted with sighted controls. The results showed increased FC between Broca's area and the visually deprived areas in the peripheral V1 for individuals with tunnel vision, and both the peripheral and central V1 for blind individuals. These findings suggest that FC can develop in the adult brain between the visual and language systems in the completely and partially blind. These changes start in the deprived areas and increase in size (involving both foveal and peripheral V1) and strength (from negative to positive FC) as the disease and sensory deprivation progress. These observations support the claim that functional connectivity between remote systems that perform completely different tasks can change in the adult brain in cases of total and even partial visual deprivation.
Murray, MM, Lewkowicz DJ, Amedi A, Wallace MT.  2016.  Multisensory Processes: A Balancing Act across the Lifespan.. Trends in neurosciences. 39(8):567-79. Abstract
Multisensory processes are fundamental in scaffolding perception, cognition, learning, and behavior. How and when stimuli from different sensory modalities are integrated rather than treated as separate entities is poorly understood. We review how the relative reliance on stimulus characteristics versus learned associations dynamically shapes multisensory processes. We illustrate the dynamism in multisensory function across two timescales: one long term that operates across the lifespan and one short term that operates during the learning of new multisensory relations. In addition, we highlight the importance of task contingencies. We conclude that these highly dynamic multisensory processes, based on the relative weighting of stimulus characteristics and learned associations, provide both stability and flexibility to brain functions over a wide range of temporal scales.
Rubin, J, Ulanovsky N, Nelken I, Tishby N.  2016.  The Representation of Prediction Error in Auditory Cortex.. PLoS computational biology. 12(8):e1005058. Abstract
To survive, organisms must extract information from the past that is relevant for their future. How this process is expressed at the neural level remains unclear. We address this problem by developing a novel approach from first principles. We show here how to generate low-complexity representations of the past that produce optimal predictions of future events. We then illustrate this framework by studying the coding of 'oddball' sequences in auditory cortex. We find that for many neurons in primary auditory cortex, trial-by-trial fluctuations of neuronal responses correlate with the theoretical prediction error calculated from the short-term past of the stimulation sequence, under constraints on the complexity of the representation of this past sequence. In some neurons, the effect of prediction error accounted for more than 50% of response variability. Reliable predictions often depended on a representation of the sequence of the last ten or more stimuli, although the representation kept only few details of that sequence.
Rubin, J, Ulanovsky N, Nelken I, Tishby N.  2016.  The Representation of Prediction Error in Auditory Cortex.. PLoS Comput Biol. 12(8): e1005058. Abstract
To survive, organisms must extract information from the past that is relevant for their future. How this process is expressed at the neural level remains unclear. We address this problem by developing a novel approach from first principles. We show here how to generate low-complexity representations of the past that produce optimal predictions of future events. We then illustrate this framework by studying the coding of 'oddball' sequences in auditory cortex. We find that for many neurons in primary auditory cortex, trial-by-trial fluctuations of neuronal responses correlate with the theoretical prediction error calculated from the short-term past of the stimulation sequence, under constraints on the complexity of the representation of this past sequence. In some neurons, the effect of prediction error accounted for more than 50% of response variability. Reliable predictions often depended on a representation of the sequence of the last ten or more stimuli, although the representation kept only few details of that sequence.
Gokhman, D, Meshorer E, Carmel L.  2016.  Epigenetics: It's Getting Old. Past Meets Future in Paleoepigenetics.. Trends in ecology & evolution. 31(4):290-300. Abstract
Recent years have witnessed the rise of ancient DNA (aDNA) technology, allowing comparative genomics to be carried out at unprecedented time resolution. While it is relatively straightforward to use aDNA to identify recent genomic changes, it is much less clear how to utilize it to study changes in epigenetic regulation. Here we review recent works demonstrating that highly degraded aDNA still contains sufficient information to allow reconstruction of epigenetic signals, including DNA methylation and nucleosome positioning maps. We discuss challenges arising from the tissue specificity of epigenetics, and show how some of them might in fact turn into advantages. Finally, we introduce a method to infer methylation states in tissues that do not tend to be preserved over time.
Pfeiffer, T, Heinze N, Frysch R, Deouell LY, Schoenfeld MA, Knight RT, Rose G.  2016.  Extracting duration information in a picture category decoding task using hidden Markov Models.. Journal of neural engineering. 13(2):026010. Abstract
Adapting classifiers for the purpose of brain signal decoding is a major challenge in brain-computer-interface (BCI) research. In a previous study we showed in principle that hidden Markov models (HMM) are a suitable alternative to the well-studied static classifiers. However, since we investigated a rather straightforward task, advantages from modeling of the signal could not be assessed.
Nitzan, E, Avraham O, Kahane N, Ofek S, Kumar D, Kalcheim C.  2016.  Dynamics of BMP and Hes1/Hairy1 signaling in the dorsal neural tube underlies the transition from neural crest to definitive roof plate.. BMC biology. 14:23. Abstract
The dorsal midline region of the neural tube that results from closure of the neural folds is generally termed the roof plate (RP). However, this domain is highly dynamic and complex, and is first transiently inhabited by prospective neural crest (NC) cells that sequentially emigrate from the neuroepithelium. It only later becomes the definitive RP, the dorsal midline cells of the spinal cord. We previously showed that at the trunk level of the axis, prospective RP progenitors originate ventral to the premigratory NC and progressively reach the dorsal midline following NC emigration. However, the molecular mechanisms underlying the end of NC production and formation of the definitive RP remain virtually unknown.
Siuda-Krzywicka, K, Bola Ł, Paplińska M, Sumera E, Jednoróg K, Marchewka A, Śliwińska MW, Amedi A, Szwed M.  2016.  Massive cortical reorganization in sighted Braille readers.. eLife. 5:e10762. Abstract
The brain is capable of large-scale reorganization in blindness or after massive injury. Such reorganization crosses the division into separate sensory cortices (visual, somatosensory...). As its result, the visual cortex of the blind becomes active during tactile Braille reading. Although the possibility of such reorganization in the normal, adult brain has been raised, definitive evidence has been lacking. Here, we demonstrate such extensive reorganization in normal, sighted adults who learned Braille while their brain activity was investigated with fMRI and transcranial magnetic stimulation (TMS). Subjects showed enhanced activity for tactile reading in the visual cortex, including the visual word form area (VWFA) that was modulated by their Braille reading speed and strengthened resting-state connectivity between visual and somatosensory cortices. Moreover, TMS disruption of VWFA activity decreased their tactile reading accuracy. Our results indicate that large-scale reorganization is a viable mechanism recruited when learning complex skills.
Cohen, E, Malka D, Shemer A, Shahmoon A, Zalevsky Z, London M.  2016.  Neural networks within multi-core optic fibers. Nature Publishing Group. Scientific Reports. 6:29080. Abstract
Hardware implementation of artificial neural networks facilitates real-time parallel processing of massive data sets. Optical neural networks offer low-volume 3D connectivity together with large bandwidth and minimal heat production in contrast to electronic implementation. Here, we present a conceptual design for in-fiber optical neural networks. Neurons and synapses are realized as individual silica cores in a multi-core fiber. Optical signals are transferred transversely between cores by means of optical coupling. Pump driven amplification in erbium-doped cores mimics synaptic interactions. We simulated three-layered feed-forward neural networks and explored their capabilities. Simulations suggest that networks can differentiate between given inputs depending on specific configurations of amplification; this implies classification and learning capabilities. Finally, we tested experimentally our basic neuronal elements using fibers, couplers, and amplifiers, and demonstrated that this configuration implements a neuron-like function. Therefore, devices similar to our proposed multi-core fiber could potentially serve as building blocks for future large-scale small-volume optical artificial neural networks.