add refs to intro

Author: qgeissmann

report/discussion.tex

@@ -66,9 +66,9 @@ \subsection{Exhaustive and time-aware feature extraction}
 This study demonstrated that addition of temporal context significantly improved predictive accuracy (fig.\ref{fig:temporal_integration}).
 The convolution approach (eq.\ref{eq:window}) appeared to provide better results.
 Instead of averaging feature after calculation, it may be advantageous to compute features over epochs of different length in a first place.
-Thus, the accuracy of local of non additive features, such as median, will be improved. In addition to local mean of feature, other variables, such as local
+Thus, the accuracy of local of non additive features, such as median, will be improved. In addition to local mean of feature, other interval variables, such as local
 slope and local variance of each feature may improve
-classification\cite{deng_time_2013}.
+classification\cite{rodriguez_support_2005,deng_time_2013}.
 
 Although addition of time-dependent variables improved accuracy over a time-unaware model, their use can be seen as controversial.
 Indeed, including prior information about sleep structure will cause problems if the aim is to find differences in sleep structure.

report/intro.tex

@@ -15,22 +15,16 @@ \section{Introduction} \label{intro}
 Classically, there are three distinct main types of sleep-related behaviours: wakefulness, \gls{nrem} sleep and \gls{rem} 
 sleep which are collectively referred to as \emph{vigilance states}.
 Vigilance states are usually defined on the basis of \gls{eeg} and \gls{emg} (fig.~\ref{fig:sleep_description}).
-When awake (WAKE), animals tend to have a high muscular activity which translates to high amplitude voltage changes in the \gls{emg}.
-____________________________________________________________________________
-[SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING]
+When awake (WAKE), animals tend to have a high muscular activity which translates to high amplitude voltage changes in the \gls{emg}\cite{crisler_sleep-stage_2008}.
 During wakefulness, \gls{eeg} is dominated by relatively low amplitude oscillations of frequencies
 between six and ten hertz often referred to as theta waves. 
-____________________________________________________________________________
-[SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING]
 In contrast, \gls{nrem} sleep, also called slow wave sleep, is a period of muscular inactivity (low \gls{emg})
 dominated by slow oscillations (below 4Hz) of high amplitude named delta waves.
-____________________________________________________________________________
-[SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING]
+\cite{vyazovskiy_nrem_2014}
 The third state, \gls{rem} sleep, is characterised by a complete lack of muscular activity (atony) and an \gls{eeg} activity very similar to the awake state.
 \gls{rem} sleep is the least prevalent of all three stages, and represents generally  20\% of all sleeping time.
 The prevalence of these three states as well as their temporal organisation are extremely important observations in sleep research
-____________________________________________________________________________
-[SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING SOURCE MISSING]
+\cite{mieda_orexin_2004,pang_unexpected_2009}
 
 \input{./figures/sleep_description}
 
@@ -92,16 +86,16 @@ \section{Introduction} \label{intro}
 forests\cite{breiman_random_2001} were, overall, the most accurate predictors.
 
 The study herein bases itself on these promising results by computing an even larger number of features.
-An important addition was the computation of time-aware
-features\cite{dietterich_machine_2002,deng_time_2013} which significantly improved
+An important addition was the computation of interval
+features (\ie{} time aware)
+\cite{dietterich_machine_2002,rodriguez_support_2005,deng_time_2013} which significantly improved
 accuracy. 
 Indeed, since manual scorers usually use contextual information to make a decision concerning a given state (personal communications),
 it was highly relevant to build a time-aware predictor.
 Furthermore, as opposed to previous work, stratified cross-validation\cite{ding_querying_2008} procedures and
 comparisons of resulting sleep patterns were performed.
 These improvements altogether contributed to achieving a very satisfying overall accuracy of 92\% 
-_____________________________________________________________________
-[IN COMPARISON TO WHAT? IN COMPARISON TO WHAT? IN COMPARISON TO WHAT?] compared to manual scoring.
+compared to manual scoring.
 
 
 In order to pave the way for the implementation of a ubiquitous sleep scoring software,

report/report.bib

@@ -1367,4 +1367,44 @@ @article{rodriguez_support_2005
 	keywords = {Boosting, Kernel methods, Time series classification},
 	pages = {171--178},
 	file = {ScienceDirect Full Text PDF:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/I9EZDS2J/Rodríguez et al. - 2005 - Support vector machines of interval-based features.pdf:application/pdf;ScienceDirect Snapshot:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/4B4QII39/S0950705105000432.html:text/html}
+}
+
+@article{pack_novel_2007,
+	title = {Novel method for high-throughput phenotyping of sleep in mice},
+	volume = {28},
+	copyright = {Copyright © 2007 the American Physiological Society},
+	issn = {1094-8341, 1531-2267},
+	url = {http://physiolgenomics.physiology.org/content/28/2/232},
+	doi = {10.1152/physiolgenomics.00139.2006},
+	abstract = {Assessment of sleep in mice currently requires initial implantation of chronic electrodes for assessment of electroencephalogram ({EEG}) and electromyogram ({EMG}) followed by time to recover from surgery. Hence, it is not ideal for high-throughput screening. To address this deficiency, a method of assessment of sleep and wakefulness in mice has been developed based on assessment of activity/inactivity either by digital video analysis or by breaking infrared beams in the mouse cage. It is based on the algorithm that any episode of continuous inactivity of ≥40 s is predicted to be sleep. The method gives excellent agreement in C57BL/6J male mice with simultaneous assessment of sleep by {EEG}/{EMG} recording. The average agreement over 8,640 10-s epochs in 24 h is 92\% (n = 7 mice) with agreement in individual mice being 88–94\%. Average {EEG}/{EMG} determined sleep per 2-h interval across the day was 59.4 min. The estimated mean difference (bias) per 2-h interval between inactivity-defined sleep and {EEG}/{EMG}-defined sleep was only 1.0 min (95\% confidence interval for mean bias −0.06 to +2.6 min). The standard deviation of differences (precision) was 7.5 min per 2-h interval with 95\% limits of agreement ranging from −13.7 to +15.7 min. Although bias significantly varied by time of day (P = 0.0007), the magnitude of time-of-day differences was not large (average bias during lights on and lights off was +5.0 and −3.0 min per 2-h interval, respectively). This method has applications in chemical mutagenesis and for studies of molecular changes in brain with sleep/wakefulness.},
+	language = {en},
+	number = {2},
+	urldate = {2014-09-08},
+	journal = {Physiological Genomics},
+	author = {Pack, Allan I. and Galante, Raymond J. and Maislin, Greg and Cater, Jacqueline and Metaxas, Dimitris and Lu, Shan and Zhang, Lin and Smith, Randy Von and Kay, Timothy and Lian, Jie and Svenson, Karen and Peters, Luanne L.},
+	month = feb,
+	year = {2007},
+	pmid = {16985007},
+	note = {Assessment of sleep in mice currently requires initial implantation of chronic electrodes for assessment of electroencephalogram ({EEG}) and electromyogram ({EMG}) followed by time to recover from surgery. Hence, it is not ideal for high-throughput screening. To address this deficiency, a method of assessment of sleep and wakefulness in mice has been developed based on assessment of activity/inactivity either by digital video analysis or by breaking infrared beams in the mouse cage. It is based on the algorithm that any episode of continuous inactivity of ≥40 s is predicted to be sleep. The method gives excellent agreement in C57BL/6J male mice with simultaneous assessment of sleep by {EEG}/{EMG} recording. The average agreement over 8,640 10-s epochs in 24 h is 92\% ( n = 7 mice) with agreement in individual mice being 88–94\%. Average {EEG}/{EMG} determined sleep per 2-h interval across the day was 59.4 min. The estimated mean difference (bias) per 2-h interval between inactivity-defined sleep and {EEG}/{EMG}-defined sleep was only 1.0 min (95\% confidence interval for mean bias −0.06 to +2.6 min). The standard deviation of differences (precision) was 7.5 min per 2-h interval with 95\% limits of agreement ranging from −13.7 to +15.7 min. Although bias significantly varied by time of day ( P = 0.0007), the magnitude of time-of-day differences was not large (average bias during lights on and lights off was +5.0 and −3.0 min per 2-h interval, respectively). This method has applications in chemical mutagenesis and for studies of molecular changes in brain with sleep/wakefulness.},
+	pages = {232--238},
+	file = {Full Text PDF:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/XMMFF5TN/Pack et al. - 2007 - Novel method for high-throughput phenotyping of sl.pdf:application/pdf;Snapshot:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/66MD7FI2/232.html:text/html}
+}
+
+@article{mieda_orexin_2004,
+	title = {Orexin peptides prevent cataplexy and improve wakefulness in an orexin neuron-ablated model of narcolepsy in mice},
+	volume = {101},
+	issn = {0027-8424, 1091-6490},
+	url = {http://www.pnas.org/content/101/13/4649},
+	doi = {10.1073/pnas.0400590101},
+	abstract = {Narcolepsy-cataplexy is a neurological disorder associated with the inability to maintain wakefulness and abnormal intrusions of rapid eye movement sleep-related phenomena into wakefulness such as cataplexy. The vast majority of narcoleptic-cataplectic individuals have low or undetectable levels of orexin (hypocretin) neuropeptides in the cerebrospinal fluid, likely due to specific loss of the hypothalamic orexin-producing neurons. Currently available treatments for narcolepsy are only palliative, symptom-oriented pharmacotherapies. Here, we demonstrate rescue of the narcolepsy-cataplexy phenotype of orexin neuron-ablated mice by genetic and pharmacological means. Ectopic expression of a prepro-orexin transgene in the brain completely prevented cataplectic arrests and other abnormalities of rapid eye movement sleep in the absence of endogenous orexin neurons. Central administration of orexin-A acutely suppressed cataplectic behavioral arrests and increased wakefulness for 3 h. These results indicate that orexin neuron-ablated mice retain the ability to respond to orexin neuropeptides and that a temporally regulated and spatially targeted secretion of orexins is not necessary to prevent narcoleptic symptoms. Orexin receptor agonists would be of potential value for treating human narcolepsy.},
+	language = {en},
+	number = {13},
+	urldate = {2014-09-08},
+	journal = {Proceedings of the National Academy of Sciences of the United States of America},
+	author = {Mieda, Michihiro and Willie, Jon T. and Hara, Junko and Sinton, Christopher M. and Sakurai, Takeshi and Yanagisawa, Masashi},
+	month = mar,
+	year = {2004},
+	pmid = {15070772},
+	pages = {4649--4654},
+	file = {Full Text PDF:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/4Q3NPZ2M/Mieda et al. - 2004 - Orexin peptides prevent cataplexy and improve wake.pdf:application/pdf;Snapshot:/home/quentin/.mozilla/firefox/kkgy4t0w.default/zotero/storage/DGBKJEEK/4649.html:text/html}
 }