Rescue of long-range circuit dysfunction in Alzheimer's disease models


Alzheimer's disease (AD) is associated with defects of synaptic connectivity. Such defects may not be restricted to local neuronal interactions but may extend to long-range brain activities, such as slow-wave oscillations that are particularly prominent during non–rapid eye movement (non-REM) sleep and are important for integration of information across distant brain regions involved in memory consolidation. There is increasing evidence that sleep is often impaired in AD, but it is unclear whether this impairment is directly related to amyloid-β (Aβ) pathology. Here we demonstrate that slow-wave activity is severely altered in the neocortex, thalamus and hippocampus in mouse models of AD amyloidosis. Most notably, our results reveal an Aβ-dependent impairment of slow-wave propagation, which causes a breakdown of the characteristic long-range coherence of slow-wave activity. The finding that the impairment can be rescued by enhancing GABAAergic inhibition identifies a synaptic mechanism underlying Aβ-dependent large-scale circuit dysfunction.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: Impaired cortical slow-wave oscillations in AD transgenic mice.
Figure 2: Breakdown of long-range coherence of cortical slow waves.
Figure 3: Functional decoupling of cortex and hippocampus.
Figure 4: Functional decoupling of cortex and thalamus.
Figure 5: Rescue of impaired slow-wave activity by GABAA-receptor agonists.
Figure 6: Exogenous amyloid-β disrupts slow-wave activity in wild-type mice.
Figure 7: Effects of low-dose clonazepam treatment on cortical slow-wave impairments as well as on learning and memory deficits in APP23 × PS45 mice.
Figure 8: An animal's performance in the water maze test correlates with the long-range coherence of the slow waves.


  1. 1

    Hardy, J. & Selkoe, D.J. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 297, 353–356 (2002).

  2. 2

    Ju, Y.-E.S., Lucey, B.P. & Holtzman, D.M. Sleep and Alzheimer disease pathology—a bidirectional relationship. Nat. Rev. Neurol. 10, 115–119 (2014).

  3. 3

    Ju, Y.E. et al. Sleep quality and preclinical Alzheimer disease. JAMA Neurol. 70, 587–593 (2013).

  4. 4

    Lim, A.S., Kowgier, M., Yu, L., Buchman, A.S. & Bennett, D.A. Sleep fragmentation and the risk of incident Alzheimer's disease and cognitive decline in older persons. Sleep 36, 1027–1032 (2013).

  5. 5

    Roh, J.H. et al. Disruption of the sleep-wake cycle and diurnal fluctuation of β-amyloid in mice with Alzheimer's disease pathology. Sci. Transl. Med. 4, 150ra122 (2012).

  6. 6

    Massimini, M., Huber, R., Ferrarelli, F., Hill, S. & Tononi, G. The sleep slow oscillation as a traveling wave. J. Neurosci. 24, 6862–6870 (2004).

  7. 7

    Chauvette, S., Volgushev, M. & Timofeev, I. Origin of active states in local neocortical networks during slow sleep oscillation. Cereb. Cortex 20, 2660–2674 (2010).

  8. 8

    Stroh, A. et al. Making waves: initiation and propagation of corticothalamic Ca2+ waves in vivo. Neuron 77, 1136–1150 (2013).

  9. 9

    Luczak, A., Bartho, P., Marguet, S.L., Buzsáki, G. & Harris, K.D. Sequential structure of neocortical spontaneous activity in vivo. Proc. Natl. Acad. Sci. USA 104, 347–352 (2007).

  10. 10

    Steriade, M., Contreras, D., Curro Dossi, R. & Nunez, A. The slow (< 1 Hz) oscillation in reticular thalamic and thalamocortical neurons: scenario of sleep rhythm generation in interacting thalamic and neocortical networks. J. Neurosci. 13, 3284–3299 (1993).

  11. 11

    Sheroziya, M. & Timofeev, I. Global intracellular slow-wave dynamics of the thalamocortical system. J. Neurosci. 34, 8875–8893 (2014).

  12. 12

    Sirota, A., Csicsvari, J., Buhl, D. & Buzsáki, G. Communication between neocortex and hippocampus during sleep in rodents. Proc. Natl. Acad. Sci. USA 100, 2065–2069 (2003).

  13. 13

    Ji, D. & Wilson, M.A. Coordinated memory replay in the visual cortex and hippocampus during sleep. Nat. Neurosci. 10, 100–107 (2007).

  14. 14

    Chauvette, S., Crochet, S., Volgushev, M. & Timofeev, I. Properties of slow oscillation during slow-wave sleep and anesthesia in cats. J. Neurosci. 31, 14998–15008 (2011).

  15. 15

    Poulet, J.F. & Petersen, C.C. Internal brain state regulates membrane potential synchrony in barrel cortex of behaving mice. Nature 454, 881–885 (2008).

  16. 16

    Hoffman, K.L. et al. The upshot of up states in the neocortex: from slow oscillations to memory formation. J. Neurosci. 27, 11838–11841 (2007).

  17. 17

    Diekelmann, S. & Born, J. The memory function of sleep. Nat. Rev. Neurosci. 11, 114–126 (2010).

  18. 18

    Tononi, G. & Cirelli, C. Sleep and the price of plasticity: from synaptic and cellular homeostasis to memory consolidation and integration. Neuron 81, 12–34 (2014).

  19. 19

    Steriade, M. Grouping of brain rhythms in corticothalamic systems. Neuroscience 137, 1087–1106 (2006).

  20. 20

    Vyazovskiy, V.V. & Harris, K.D. Sleep and the single neuron: the role of global slow oscillations in individual cell rest. Nat. Rev. Neurosci. 14, 443–451 (2013).

  21. 21

    Sperling, R.A. et al. Functional alterations in memory networks in early Alzheimer's disease. Neuromolecular Med. 12, 27–43 (2010).

  22. 22

    Bero, A.W. et al. Bidirectional relationship between functional connectivity and amyloid-β deposition in mouse brain. J. Neurosci. 32, 4334–4340 (2012).

  23. 23

    Raichle, M.E. Two views of brain function. Trends Cogn. Sci. 14, 180–190 (2010).

  24. 24

    Busche, M.A. et al. Clusters of hyperactive neurons near amyloid plaques in a mouse model of Alzheimer's disease. Science 321, 1686–1689 (2008).

  25. 25

    Busche, M.A. et al. Critical role of soluble amyloid-β for early hippocampal hyperactivity in a mouse model of Alzheimer's disease. Proc. Natl. Acad. Sci. USA 109, 8740–8745 (2012).

  26. 26

    Grienberger, C. et al. Staged decline of neuronal function in vivo in an animal model of Alzheimer's disease. Nat. Commun. 3, 774 (2012).

  27. 27

    Eichhoff, G., Busche, M.A. & Garaschuk, O. In vivo calcium imaging of the aging and diseased brain. Eur. J. Nucl. Med. Mol. Imaging 35 (suppl. 1): S99–S106 (2008).

  28. 28

    Stosiek, C., Garaschuk, O., Holthoff, K. & Konnerth, A. In vivo two-photon calcium imaging of neuronal networks. Proc. Natl. Acad. Sci. USA 100, 7319–7324 (2003).

  29. 29

    Grienberger, C. et al. Sound-evoked network calcium transients in mouse auditory cortex in vivo. J. Physiol. (Lond.) 590, 899–918 (2012).

  30. 30

    Kajikawa, Y. & Schroeder, C.E. How local is the local field potential? Neuron 72, 847–858 (2011).

  31. 31

    Rochefort, N.L. et al. Sparsification of neuronal activity in the visual cortex at eye-opening. Proc. Natl. Acad. Sci. USA 106, 15049–15054 (2009).

  32. 32

    Mohajerani, M.H., McVea, D.A., Fingas, M. & Murphy, T.H. Mirrored bilateral slow-wave cortical activity within local circuits revealed by fast bihemispheric voltage-sensitive dye imaging in anesthetized and awake mice. J. Neurosci. 30, 3745–3751 (2010).

  33. 33

    Hahn, T.T., Sakmann, B. & Mehta, M.R. Differential responses of hippocampal subfields to cortical up-down states. Proc. Natl. Acad. Sci. USA 104, 5169–5174 (2007).

  34. 34

    Crunelli, V. & Hughes, S.W. The slow (<1 Hz) rhythm of non-REM sleep: a dialogue between three cardinal oscillators. Nat. Neurosci. 13, 9–17 (2010).

  35. 35

    Verret, L. et al. Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model. Cell 149, 708–721 (2012).

  36. 36

    Sanchez-Vives, M.V. & McCormick, D.A. Cellular and network mechanisms of rhythmic recurrent activity in neocortex. Nat. Neurosci. 3, 1027–1034 (2000).

  37. 37

    Sigel, E. & Buhr, A. The benzodiazepine binding site of GABAA receptors. Trends Pharmacol. Sci. 18, 425–429 (1997).

  38. 38

    Selkoe, D.J. Resolving controversies on the path to Alzheimer's therapeutics. Nat. Med. 17, 1060–1065 (2011).

  39. 39

    Phillips, K.G. et al. Decoupling of sleep-dependent cortical and hippocampal interactions in a neurodevelopmental model of schizophrenia. Neuron 76, 526–533 (2012).

  40. 40

    Busche, M.A. & Konnerth, A. Neuronal hyperactivity – a key defect in Alzheimer's disease? Bioessays 37, 624–632 (2015).

  41. 41

    Cirrito, J.R. et al. Synaptic activity regulates interstitial fluid amyloid-beta levels in vivo. Neuron 48, 913–922 (2005).

  42. 42

    Huang, Y. et al. Effects of age and amyloid deposition on Aβ dynamics in the human central nervous system. Arch. Neurol. 69, 51–58 (2012).

  43. 43

    Tampellini, D. et al. Effects of synaptic modulation on β-amyloid, synaptophysin, and memory performance in Alzheimer's disease transgenic mice. J. Neurosci. 30, 14299–14304 (2010).

  44. 44

    Tannenbaum, C., Paquette, A., Hilmer, S., Holroyd-Leduc, J. & Carnahan, R. A systematic review of amnestic and non-amnestic mild cognitive impairment induced by anticholinergic, antihistamine, GABAergic and opioid drugs. Drugs Aging 29, 639–658 (2012).

  45. 45

    Billioti de Gage, S. et al. Benzodiazepine use and risk of Alzheimer's disease: case-control study. Br. Med. J. 349, g5205 (2014).

  46. 46

    Fastbom, J., Forsell, Y. & Winblad, B. Benzodiazepines may have protective effects against Alzheimer disease. Alzheimer Dis. Assoc. Disord. 12, 14–17 (1998).

  47. 47

    Genzel, L., Kroes, M.C., Dresler, M. & Battaglia, F.P. Light sleep versus slow wave sleep in memory consolidation: a question of global versus local processes? Trends Neurosci. 37, 10–19 (2014).

  48. 48

    Andrews-Zwilling, Y. et al. Apolipoprotein E4 causes age- and tau-dependent impairment of GABAergic interneurons, leading to learning and memory deficits in mice. J. Neurosci. 30, 13707–13717 (2010).

  49. 49

    Achermann, P. & Borbely, A.A. Low-frequency (< 1 Hz) oscillations in the human sleep electroencephalogram. Neuroscience 81, 213–222 (1997).

  50. 50

    Buzsáki, G., Logothetis, N. & Singer, W. Scaling brain size, keeping timing: evolutionary preservation of brain rhythms. Neuron 80, 751–764 (2013).

  51. 51

    de Cheveigné, A., Edeline, J.M., Gaucher, Q. & Gourevitch, B. Component analysis reveals sharp tuning of the local field potential in the guinea pig auditory cortex. J. Neurophysiol. 109, 261–272 (2013).

  52. 52

    Paxinos, G. & Franklin, K.B.J. The Mouse Brain in Stereotaxic Coordinates (Academic, 2001).

  53. 53

    Castro-Alamancos, M.A. Absence of rapid sensory adaptation in neocortex during information processing states. Neuron 41, 455–464 (2004).

Download references


We thank C. Tischbirek for assistance with graphics, A. Beazley for help with data analysis and M. Staufenbiel for discussions and comments. This work was funded by an Advanced European Research Council grant to A.K., the European Union FP7 program (Project Corticonic) and the Deutsche Forschungsgemeinschaft (RTG 1373 and SFB870). M.A.B. was supported by the Langmatz Stiftung. I.N. was supported by a grant from the Israel Science Foundation and by the Deutsche Forschungsgemeinschaft (SFB870).

Author information




M.A.B. and A.K. designed the study; M.A.B., M.K., H.A. and T.N. performed the experiments; M.A.B., M.K., H.A., H.F., I.N. and A.K. performed the analysis; M.A.B., I.N. and A.K. wrote the manuscript.

Corresponding authors

Correspondence to Marc Aurel Busche or Arthur Konnerth.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Large-scale calcium imaging in vivo.

(a) Schematic illustration of the in vivo camera imaging set-up for large-scale calcium imaging. (b) High-magnification two-photon image from a region in layer 2/3 of the frontal cortex and representative activity traces from neurons (1 – 3), neuropil and the entire image frame (integral).

Supplementary Figure 2 Correlations across brain regions.

(a) Schematic representation of the cortical area that was stained with fluorescent calcium indicator OGB-1 AM (left panel). The cortical subregions were identified by using a mouse brain atlas (right panel). Note that midline structures (cingulate and retrosplenial cortex) were not labeled by the calcium indicator. (b) Regions of interest for cortical imaging. (c) Regions for simultaneous recordings from cortex and contra-lateral hippocampus. (d) Regions for simultaneous recordings from cortex and contra-lateral thalamus.

Supplementary Figure 3 Average correlation matrix from all calcium imaging experiments.

Mean correlation coefficient plotted for all pairs of cortical areas in wild-type (WT, black) and APP23 x PS45 (Tg, red) mice. Occ. = occipital, som. = somatosensory, mot. = motor and fro. = frontal cortex. Error bars correspond to one standard error of the mean. Note that pairs of areas that have the same distance from each other (e.g. occipital and somatosensory vs. motor and frontal, both being adjacent pairs, or occipital and motor vs. somatosensory and frontal, both pairs having one intervening area between them) also have about the same correlation values (the correlation matrices have approximately constant diagonals).

Supplementary Figure 4 Impairment of the correlation between brain areas occurs predominantly in the slow wave and delta range, but is present in all frequency bands.

(a-d) The mean correlation coefficients of the fluorescent calcium signals are plotted as a function of cortical distance in wild-type (WT, black) and APP23 x PS45 (Tg, red) mice. Panel a presents the same data as in Supplementary Fig. 3 (the correlations calculated from the broadband signal). Instead of plotting the data in 3-d, the correlation coefficients from the four areas (six coefficients in total) are plotted separated by the cortical distance between the areas (these are the diagonal of the correlation matrix), demonstrating the dependence of the correlation on cortical distance. Near = two nearby cortical domains (e.g., occipital and somatosensory cortex), mid = two domains separated by one region (e.g., occipital and motor cortex), far = two domains separated by two regions (always occipital and frontal cortex). Panels b-d show the correlations, plotted in the same way, for the different frequency bands. The slow wave and delta correlation values (0.1 – 3 Hz, b) are essentially the same as the broadband (0.1 – 30 Hz, a) values, because this is the frequency range with the highest power (see also Supplementary Fig. 12). However, in all frequency bands, the correlations are higher in the WT than in the Tg animals.

Supplementary Figure 5 Impairment of correlation between brain areas is reflected in phase consistency in all frequency bands.

The average pairwise phase consistency (PPC; Vinck, M., et al. (2010) NeuroImage 51, 112-122) is plotted in the format as in Supplementary Fig. 4. The PPC was calculated for 32 s long segments of calcium recording, as elsewhere in the paper, using a multitaper method (Mitra, P.P. & Pesaran, B. (1999) Biophys. J. 76, 691-708) with NW = 16 (32 tapers). The PPC was averaged over frequency bands in order to be comparable to the correlation coefficients displayed in the other figures. For all anatomical distances and all frequency bands, the PPC in the WT animals (black) is larger than in the Tg animals (red). The PPC was analyzed using a linear mixed effect model with anatomical distance and genetic background as fixed factors, and with mouse and hemisphere within mouse as random factors. The effect of transgenic background was highly significant in all frequency bands. Effect of genetic background on PPC in the slow-wave frequency range was highly significant: F(1,450) = 65.8, P = 4.8e-15. In the theta range, the effect of genetic background was not significant. In the alpha range, there was a significant interaction between genetic background and anatomical distance, with phase consistency decreasing faster with distance in the Tg animals: F(2,450) = 6.09, P = 0.0025. The lack of significance of genetic background in the theta band, and its presence in the alpha band, were also found for LFP recordings in unanesthetized animals (see Supplementary Fig. 18).

Supplementary Figure 6 Similar impairment of long-range coherence of slow waves in APP23 × PS45 and APP23 mice.

(a) Summary graph displaying the average cross-correlation coefficients and standard errors plotted against the cortical distance (near = two nearby cortical domains, mid = two cortical domains separated by one region, far = two domains separated by two regions) in wild-type (WT, black, n = 9), APP23 (blue, n = 3) and APP23 x PS45 (red, n = 9) mice, respectively. (b) Bar graph shows the average correlation coefficients between cortical and hippocampal activity for WT (n = 5), APP23 (n = 3) and APP23 x PS45 (n = 5) mice, respectively. WT and APP23 x PS45: two-sample t-test, t(108) = 17, P = 2.7e-32 (these are the same data as shown in Fig. 3i of the main text); WT and APP23: t(75) = 21.6, P = 5.7e-34. Error bars denote SEM.

Supplementary Figure 7 LFP recordings in anesthetized animals yielded results that were similar to those obtained with calcium fluorescence imaging.

(a) Scheme illustrating the positions of the LFP electrodes. (b) Examples of LFP recordings in a wild-type (WT, top panel) and an APP23 x PS45 (Tg, bottom panel) mouse. Superimposed traces are from the shaded areas. (c) Bar graph shows the mean correlation coefficient between occipital and frontal cortex (ctx-ctx), frontal cortex and hippocampus (ctx-hc) as well as frontal cortex and thalamus (ctx-thl) for WT (gray bars) and Tg (green bars) mice (n = 18 animals), respectively. Two-way ANOVA on genotype (WT vs. Tg) and correlation type (ctx-ctx, ctx-hc and ctx-thl): main effect of genotype F(1,14) = 197, P = 1.2e-9; main effect of correlation type: F(2,14) = 0.27, P = 0.77; interaction: F(2,14) = 0.09, P = 0.91. Thus, all three correlation types are equally depressed in Tg relative to WT mice. Error bars denote SEM.

Supplementary Figure 8 Correlation between calcium transients and local field potentials in the hippocampus.

Simultaneous recording of the local field potential (LFP was inverted, blue) and camera-fluorescence signals (black) in the dorsal hippocampus.

Supplementary Figure 9 Correlation between calcium transients and local field potentials in the thalamus.

Simultaneous recording of the local field potential (LFP was inverted, blue) and camera-fluorescence signals (black) in the thalamus.

Supplementary Figure 10 Cortico-thalamic correlation maps.

Correlation maps using a seed-pixel in cortical regions (top and middle) or in the thalamus (bottom) in a wild-type (WT, left panel) and an APP23 x PS45 (Tg, right panel) mouse, respectively. The white asterisks denote the location of the seed pixel.

Supplementary Figure 11 Histological identification of local field potential electrode position.

(a, b) Example sections showing the electrode location in the hippocampus (a, top panel) and in the thalamus (b, bottom panel). Red: DiI-labeled electrode track. White arrows indicate the inferred location of the electrode tip.

Supplementary Figure 12 Power spectra of calcium imaging data.

Power spectra of cortical calcium fluorescence time series in wild-type (WT) and APP23 x PS45 (Tg) mice. While all power spectra show clear peaks at 0.3 Hz, the peak bandwidth increased substantially in the Tg mice, reflecting the higher rates of calcium transients in the Tg animals (see Fig. 2h of the main text). Note that, in the Tg mice, there is a greater increase in bandwidth in the frontal cortex (solid red line) compared to the occipital cortex (solid blue line). This finding is consistent with the relatively larger increase in the calcium wave rate in these areas (see Fig. 2h of the main text).

Supplementary Figure 13 The glutamate receptor antagonist APV has no effect on impaired slow-wave activity in APP23 × PS45 mice.

Superimposed traces of slow-wave activity in the frontal (red) and the occipital (black) cortex of an APP23 x PS45 mouse before (control), during and after (wash-out) application of 50 µM APV.

Supplementary Figure 14 Correlation maps of the pharmacological experiments.

(a-d) Correlation maps of the pharmacological experiments shown in Figs. 5 and 6 of the main text (topical midazolam application in Fig. 5a, topical gabazine application in Fig. 5b, topical Aβ application in Fig. 6a and rescue of topical Aβ application with midazolam in Fig. 6b). White asterisks mark the seed pixel that was used for the generation of the maps.

Supplementary Figure 15 Aβ1–40 and Aβ1–42 lead to similar impairments of coherent activity.

(a, b) Summary graph showing the average cross-correlation coefficients and standard errors plotted against the cortical distance (near = two nearby cortical domains, mid = two cortical domains separated by one region, far = two domains separated by two regions) before (black), during (red) and after (gray) wash-in of amyloid-β (Aβ) 1-40 in a or Aβ 1-42 in b in WT mice (n = 6 animals). For statistics see Supplementary Table 1.

Supplementary Figure 16 Determination of sleep and wake states in unanesthetized animals.

(a, b) Representative traces of surface EMG (top), cortical LFP (middle) and power spectral density of the LFP (bottom) under sleep (left) and wake (right) states in a wild-type (WT, a) and APP23 x PS45 (Tg, b) mouse, respectively. Sleep was defined as the prolonged periods with immobility of the animal, absence of EMG activity, closed eyes, and large-amplitude, low frequency activity in the LFP recordings. Wakefulness was characterized by periods with high EMG activity, occasional locomotion, opened eyes, as well as lower amplitude and higher frequency LFP activity. (c,d) Median power spectra of 200-sec LFP traces during sleep (left) and waking (right) in WT (c) and Tg (d) mice (n = 8 animals). Shaded error bands represent upper and lower quartile of the power spectra.

Supplementary Figure 17 Impairment of long-range coherence of slow-wave activity in naturally sleeping APP23 × PS45 mice.

(a, b) Example traces of LFP recordings from the frontal (red) and the occipital (black) cortex in unanesthetized wild-type (WT, a) and APP23 x PS45 (Tg, b) mice, respectively. (c) Example traces of LFP recordings from the frontal (red) and occipital (black) cortex in a Tg mouse before and 10 min after intraperitoneal injection of a low-dose of clonazepam. (d) Bar graph shows that the mean correlation between frontal and occipital cortex in the slow wave and delta range was significantly lower in Tg compared to WT mice (n = 10 animals; t(10) = 2.69, P = 0.023, two-sample t-test), and that, in the Tg mice, treatment with clonazepam increased this correlation (t(5) = 4.83, P = 0.0048, two-sample t-test).

Supplementary Figure 18 Impairment of long-range coherence in non-anesthetized animals extends to higher frequency bands.

Bar graph displays the mean correlation coefficients between LFP recordings from frontal and occipital cortex at different frequency bands in naturally sleeping wild-type (WT) and APP23 x PS45 (Tg) mice. Note that in the Tg mice (red bars) the correlation coefficients are considerably smaller compared to WT mice (gray bars), and that a low-dose of clonazepam (blue bars) increased the correlations. For the differences between WT and Tg, one-way nested ANOVA with animals nested within genotype, F(1,3) = 40.7, 11.6, 7.23, 13.2, 10.2 (P = 0.0078, 0.042, 0.074, 0.036, 0.049) for the broadband, slow wave and delta, theta, alpha and beta ranges. For the differences before and after clonazepam, two-sample t-test, t(5) = 4.07, 4.83, 0.35, 3.82, 4.30 (P = 0.0096, 0.0048, 0.74, 0.012, 0.0077) in the same ranges. The data in the slow wave and delta range is the same as in Supplementary Fig. 17. Error bars indicate SEM.

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–18 and Supplementary Table 1 (PDF 2104 kb)

Supplementary Methods Checklist (PDF 421 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Busche, M., Kekuš, M., Adelsberger, H. et al. Rescue of long-range circuit dysfunction in Alzheimer's disease models. Nat Neurosci 18, 1623–1630 (2015).

Download citation

Further reading