c8b19c8928f17be749658fdf1db06733 MAT_a2022v10p2786.pdf 192a012e7ba333fc18701afac59c243bf9c944c4 MAT_a2022v10p2786.pdf 4283f45f670844c187e8e667f543b79589a5a4f2096b7836206ebb2c198b54a4 MAT_a2022v10p2786.pdf Title: Detection of Abnormal Cardiac Response Patterns in Cardiac Tissue Using Deep Learning Subject: This study reports a method for the detection of mechanical signaling anomalies in cardiac tissue through the use of deep learning and the design of two anomaly detectors. In contrast to anomaly classifiers, anomaly detectors allow accurate identification of the time position of the anomaly. The first detector used a recurrent neural network (RNN) of long short-term memory (LSTM) type, while the second used an autoencoder. Mechanical contraction data present several challanges, including high presence of noise due to the biological variability in the contraction response, noise introduced by the data acquisition chain and a wide variety of anomalies. Therefore, we present a robust deep-learning-based anomaly detection framework that addresses these main issues, which are difficult to address with standard unsupervised learning techniques. For the time series recording, an experimental model was designed in which signals of cardiac mechanical contraction (right and left atria) of a CD-1 mouse could be acquired in an automatic organ bath, reproducing the physiological conditions. In order to train the anomaly detection models and validate their performance, a database of synthetic signals was designed (n = 800 signals), including a wide range of anomalous events observed in the experimental recordings. The detector based on the LSTM neural network was the most accurate. The performance of this detector was assessed by means of experimental mechanical recordings of cardiac tissue of the right and left atria. Keywords: deep learning; autoencoder; cardiac tissue; electrophysiology; electrostimulation; anomaly detection; recurrent neural network; long short-term memory; CD-1 mouse model Author: Xavier Marimon, Sara Traserra, Marcel Jiménez, Andrés Ospina and Raúl Benítez Creator: LaTeX with hyperref Producer: pdfTeX-1.40.21 CreationDate: Tue Aug 9 09:31:18 2022 CEST ModDate: Thu Aug 11 13:23:52 2022 CEST Tagged: no UserProperties: no Suspects: no Form: none JavaScript: no Pages: 21 Encrypted: no Page size: 595.276 x 841.89 pts (A4) Page rot: 0 File size: 4023502 bytes Optimized: no PDF version: 1.7 name type encoding emb sub uni object ID ------------------------------------ ----------------- ---------------- --- --- --- --------- LWJOET+VnURWPalladioL Type 1 Custom yes yes no 10 0 QQOYXG+URWPalladioL-Bold Type 1 Custom yes yes no 15 0 FRJPFW+URWPalladioL-Roma Type 1 Custom yes yes no 20 0 EFNNUC+URWPalladioL-Ital Type 1 Custom yes yes no 24 0 NBDQEL+VnURWPalladioL-Bold Type 1 Custom yes yes no 28 0 DVTHKG+TimesNewRomanPS-BoldItalicMT Type 1C WinAnsi yes yes no 46 0 WARZRX+CMSY10 Type 1 Builtin yes yes no 71 0 EIMZQL+PazoMath Type 1 Builtin yes yes no 78 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 85 0 CFDDHB+PalatinoLinotype CID TrueType Identity-H yes yes yes 88 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 94 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 97 0 GALMHL+PazoMath-Italic Type 1 Builtin yes yes no 104 0 RCDSLW+URWPalladioL-BoldItal Type 1 Custom yes yes no 108 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 118 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 121 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 124 0 CFDHFP+Calibri CID TrueType Identity-H yes yes yes 127 0 CCWEJM+CMEX10 Type 1 Builtin yes yes no 174 0 SOSTRQ+CMR10 Type 1 Builtin yes yes no 178 0 CFDHKO+Symbol Type 1C Custom yes yes yes 185 0 CFDHLP+Courier Type 1C WinAnsi yes yes no 190 0 CFDHMA+Courier-Oblique Type 1C WinAnsi yes yes no 193 0 CFDHLO+Arial CID TrueType Identity-H yes yes yes 196 0 CFDHNB+Arial,Bold CID TrueType Identity-H yes yes yes 202 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 208 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 211 0 CFDDHD+PalatinoLinotype,Italic CID TrueType Identity-H yes yes yes 214 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 220 0 CFDHKO+Symbol Type 1C Custom yes yes yes 229 0 CFDHLP+Courier Type 1C WinAnsi yes yes no 234 0 CFDIBB+Symbol CID TrueType Identity-H yes yes yes 237 0 CFDICB+MTExtra CID TrueType Identity-H yes yes yes 243 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 249 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 252 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 255 0 CFDHKO+Symbol Type 1C Custom yes yes yes 265 0 CFDHLP+Courier Type 1C WinAnsi yes yes no 270 0 CFDHMA+Courier-Oblique Type 1C WinAnsi yes yes no 273 0 CFDICB+MTExtra CID TrueType Identity-H yes yes yes 276 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 282 0 CFDDHB+PalatinoLinotype CID TrueType Identity-H yes yes yes 285 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 291 0 CFDDHD+PalatinoLinotype,Italic CID TrueType Identity-H yes yes yes 294 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 300 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 310 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 313 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 316 0 CFDHKO+Symbol Type 1C Custom yes yes yes 326 0 CFDKLD+Courier-BoldOblique Type 1C WinAnsi yes yes no 331 0 CFDKLE+Wingdings CID TrueType Identity-H yes yes yes 334 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 340 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 343 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 346 0 CFDHKO+Symbol Type 1C Custom yes yes yes 361 0 CFDHLP+Courier Type 1C WinAnsi yes yes no 366 0 CFDHMA+Courier-Oblique Type 1C WinAnsi yes yes no 369 0 CFDICB+MTExtra CID TrueType Identity-H yes yes yes 372 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 378 0 CFDDHB+PalatinoLinotype CID TrueType Identity-H yes yes yes 381 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 387 0 CFDDHD+PalatinoLinotype,Italic CID TrueType Identity-H yes yes yes 390 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 396 0 BYXUKI+EURM10 Type 1 Builtin yes yes no 409 0 CFDHKO+Symbol Type 1C Custom yes yes yes 416 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 421 0 CFDDHB+PalatinoLinotype CID TrueType Identity-H yes yes yes 424 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 430 0 CFDDHD+PalatinoLinotype,Italic CID TrueType Identity-H yes yes yes 433 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 439 0 CFDHKO+Symbol Type 1C Custom yes yes yes 449 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 454 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 457 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 460 0 CFDOCB+Calibri,Bold CID TrueType Identity-H yes yes yes 471 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 477 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 480 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 483 0 CFDHNB+Arial,Bold CID TrueType Identity-H yes yes yes 517 0 CFDDHA+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 523 0 CFDDHC+PalatinoLinotype TrueType WinAnsi yes yes no 526 0 CFDDHE+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 529 0 CFDHFP+Calibri CID TrueType Identity-H yes yes yes 532 0 Jhove (Rel. 1.22.1, 2019-04-17) Date: 2023-03-08 02:42:28 CET RepresentationInformation: MAT_a2022v10p2786.pdf ReportingModule: PDF-hul, Rel. 1.12.1 (2019-04-17) LastModified: 2023-03-07 14:05:24 CET Size: 4023502 Format: PDF Version: 1.7 Status: Well-Formed and valid SignatureMatches: PDF-hul MIMEtype: application/pdf PDFMetadata: Objects: 746 FreeObjects: 1 IncrementalUpdates: 0 DocumentCatalog: PageLayout: SinglePage PageMode: UseNone Outlines: Item: Title: Introduction Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@44c8afef Item: Title: Materials and Methods Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@7b69c6ba Children: Item: Title: Preparation of Atrial Tissue Samples Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@46daef40 Children: Item: Title: Mouse Preparation Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@12f41634 Item: Title: Atria Isolation Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@13c27452 Item: Title: Mechanical Recordings Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@262b2c86 Item: Title: Synthetic Dataset Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@371a67ec Item: Title: Architecture of Anomaly Detectors Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@5ed828d Children: Item: Title: Detector Based on an LSTM Network (D.1) Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@50d0686 Item: Title: Autoencoder-Based Detector (D.2) Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@7a3d45bd Item: Title: Results Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@1e7c7811 Children: Item: Title: Synthetic Data Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@77ec78b9 Item: Title: Experimental Data Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@1a3869f4 Item: Title: Discussion Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@a38d7a3 Item: Title: References Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@77f99a05 Info: Title: Detection of Abnormal Cardiac Response Patterns in Cardiac Tissue Using Deep Learning Author: Xavier Marimon, Sara Traserra, Marcel Jiménez, Andrés Ospina and Raúl Benítez Subject: This study reports a method for the detection of mechanical signaling anomalies in cardiac tissue through the use of deep learning and the design of two anomaly detectors. In contrast to anomaly classifiers, anomaly detectors allow accurate identification of the time position of the anomaly. The first detector used a recurrent neural network (RNN) of long short-term memory (LSTM) type, while the second used an autoencoder. Mechanical contraction data present several challanges, including high presence of noise due to the biological variability in the contraction response, noise introduced by the data acquisition chain and a wide variety of anomalies. Therefore, we present a robust deep-learning-based anomaly detection framework that addresses these main issues, which are difficult to address with standard unsupervised learning techniques. For the time series recording, an experimental model was designed in which signals of cardiac mechanical contraction (right and left atria) of a CD-1 mouse could be acquired in an automatic organ bath, reproducing the physiological conditions. In order to train the anomaly detection models and validate their performance, a database of synthetic signals was designed (n = 800 signals), including a wide range of anomalous events observed in the experimental recordings. The detector based on the LSTM neural network was the most accurate. The performance of this detector was assessed by means of experimental mechanical recordings of cardiac tissue of the right and left atria. Keywords: deep learning; autoencoder; cardiac tissue; electrophysiology; electrostimulation; anomaly detection; recurrent neural network; long short-term memory; CD-1 mouse model Creator: LaTeX with hyperref Producer: pdfTeX-1.40.21 CreationDate: Tue Aug 09 09:31:18 CEST 2022 ModDate: Thu Aug 11 13:23:52 CEST 2022 ID: 0xa0e1d92c1bb67f689454b51b6862dfbe, 0xa0e1d92c1bb67f689454b51b6862dfbe Filters: FilterPipeline: FlateDecode Fonts: Type1: Font: BaseFont: SOSTRQ+CMR10 FontSubset: true FirstChar: 40 LastChar: 93 FontDescriptor: FontName: SOSTRQ+CMR10 Flags: Symbolic FontBBox: -40, -250, 1009, 750 FontFile: true Font: BaseFont: FRJPFW+URWPalladioL-Roma FontSubset: true FirstChar: 2 LastChar: 248 FontDescriptor: FontName: FRJPFW+URWPalladioL-Roma Flags: Symbolic FontBBox: -166, -283, 1021, 943 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: WARZRX+CMSY10 FontSubset: true FirstChar: 0 LastChar: 114 FontDescriptor: FontName: WARZRX+CMSY10 Flags: Symbolic FontBBox: -29, -960, 1116, 775 FontFile: true Font: BaseFont: EFNNUC+URWPalladioL-Ital FontSubset: true FirstChar: 44 LastChar: 121 FontDescriptor: FontName: EFNNUC+URWPalladioL-Ital Flags: Symbolic FontBBox: -170, -305, 1010, 941 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: GALMHL+PazoMath-Italic FontSubset: true FirstChar: 35 LastChar: 122 FontDescriptor: FontName: GALMHL+PazoMath-Italic Flags: Symbolic FontBBox: -70, -277, 902, 733 FontFile: true Font: BaseFont: BYXUKI+EURM10 FontSubset: true FirstChar: 21 LastChar: 21 FontDescriptor: FontName: BYXUKI+EURM10 Flags: Symbolic FontBBox: -32, -243, 1060, 720 FontFile: true Font: BaseFont: LWJOET+VnURWPalladioL FontSubset: true FirstChar: 177 LastChar: 226 FontDescriptor: FontName: LWJOET+VnURWPalladioL Flags: Symbolic FontBBox: -166, -283, 1021, 943 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: NBDQEL+VnURWPalladioL-Bold FontSubset: true FirstChar: 178 LastChar: 243 FontDescriptor: FontName: NBDQEL+VnURWPalladioL-Bold Flags: Symbolic FontBBox: -152, -301, 1000, 935 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: RCDSLW+URWPalladioL-BoldItal FontSubset: true FirstChar: 78 LastChar: 114 FontDescriptor: FontName: RCDSLW+URWPalladioL-BoldItal Flags: Symbolic FontBBox: -170, -300, 1073, 935 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: EIMZQL+PazoMath FontSubset: true FirstChar: 87 LastChar: 229 FontDescriptor: FontName: EIMZQL+PazoMath Flags: Symbolic FontBBox: -40, -283, 878, 946 FontFile: true Font: BaseFont: CCWEJM+CMEX10 FontSubset: true FirstChar: 50 LastChar: 101 FontDescriptor: FontName: CCWEJM+CMEX10 Flags: Symbolic FontBBox: -24, -2960, 1454, 772 FontFile: true Font: BaseFont: QQOYXG+URWPalladioL-Bold FontSubset: true FirstChar: 3 LastChar: 248 FontDescriptor: FontName: QQOYXG+URWPalladioL-Bold Flags: Symbolic FontBBox: -152, -301, 1000, 935 FontFile: true EncodingDictionary: Differences: true Pages: Page: Sequence: 1 Annotations: Annotation: Subtype: Link Rect: 82, 323, 152, 332 Annotation: Subtype: Link Rect: 34, 312, 107, 320 Annotation: Subtype: Link Rect: 34, 164, 96, 186 Annotation: Subtype: Link Rect: 126, 90, 154, 99 Annotation: Subtype: Link Rect: 34, 78, 147, 87 Annotation: Subtype: Link Rect: 34, 756, 173, 792 Annotation: Subtype: Link Rect: 515, 756, 560, 792 Annotation: Subtype: Link Rect: 135, 660, 146, 674 Annotation: Subtype: Link Rect: 298, 660, 309, 674 Annotation: Subtype: Link Rect: 482, 660, 493, 674 Annotation: Subtype: Link Rect: 284, 181, 291, 189 ActionDest: 593 Annotation: Subtype: Link Rect: 496, 130, 503, 139 ActionDest: 593 Annotation: Subtype: Link Rect: 185, 105, 192, 114 ActionDest: 593 Annotation: Subtype: Link Rect: 495, 93, 502, 102 ActionDest: 593 Annotation: Subtype: Link Rect: 309, 67, 316, 76 ActionDest: 593 Annotation: Subtype: Link Rect: 130, 17, 273, 28 Annotation: Subtype: Link Rect: 391, 17, 560, 28 Annotation: Subtype: Link Rect: 34, 396, 87, 415 Page: Sequence: 2 Annotations: Annotation: Subtype: Link Rect: 245, 723, 252, 732 ActionDest: 593 Annotation: Subtype: Link Rect: 327, 547, 334, 556 ActionDest: 593 Annotation: Subtype: Link Rect: 411, 535, 418, 544 ActionDest: 593 Annotation: Subtype: Link Rect: 316, 459, 322, 468 ActionDest: 593 Annotation: Subtype: Link Rect: 326, 459, 337, 468 ActionDest: 596 Annotation: Subtype: Link Rect: 358, 409, 365, 418 ActionDest: 593 Annotation: Subtype: Link Rect: 200, 384, 212, 393 ActionDest: 593 Annotation: Subtype: Link Rect: 212, 384, 224, 393 ActionDest: 593 Annotation: Subtype: Link Rect: 337, 372, 349, 380 ActionDest: 593 Annotation: Subtype: Link Rect: 349, 371, 361, 380 ActionDest: 596 Annotation: Subtype: Link Rect: 531, 372, 542, 380 ActionDest: 596 Annotation: Subtype: Link Rect: 190, 346, 201, 355 ActionDest: 596 Annotation: Subtype: Link Rect: 358, 346, 369, 355 ActionDest: 596 Annotation: Subtype: Link Rect: 544, 346, 556, 355 ActionDest: 596 Annotation: Subtype: Link Rect: 302, 334, 314, 343 ActionDest: 596 Annotation: Subtype: Link Rect: 317, 334, 329, 343 ActionDest: 596 Annotation: Subtype: Link Rect: 456, 334, 468, 343 ActionDest: 596 Annotation: Subtype: Link Rect: 468, 334, 480, 343 ActionDest: 596 Annotation: Subtype: Link Rect: 401, 196, 413, 205 ActionDest: 596 Annotation: Subtype: Link Rect: 204, 183, 216, 192 ActionDest: 596 Annotation: Subtype: Link Rect: 219, 183, 231, 192 ActionDest: 596 Annotation: Subtype: Link Rect: 415, 170, 427, 180 ActionDest: 596 Annotation: Subtype: Link Rect: 343, 120, 355, 129 ActionDest: 596 Page: Sequence: 3 Annotations: Annotation: Subtype: Link Rect: 293, 620, 300, 632 ActionDest: 68 Annotation: Subtype: Link Rect: 203, 607, 218, 619 ActionDest: 68 Annotation: Subtype: Link Rect: 231, 595, 245, 607 ActionDest: 75 Annotation: Subtype: Link Rect: 389, 582, 404, 594 ActionDest: 101 Annotation: Subtype: Link Rect: 454, 570, 469, 582 ActionDest: 101 Annotation: Subtype: Link Rect: 551, 570, 558, 582 ActionDest: 406 Annotation: Subtype: Link Rect: 291, 544, 306, 556 ActionDest: 406 Annotation: Subtype: Link Rect: 553, 532, 560, 544 ActionDest: 511 Annotation: Subtype: Link Rect: 547, 342, 554, 354 ActionDest: 68 Annotation: Subtype: Link Rect: 333, 112, 340, 124 ActionDest: 75 Page: Sequence: 4 Annotations: Annotation: Subtype: Link Rect: 352, 550, 359, 562 ActionDest: 75 Page: Sequence: 5 Annotations: Annotation: Subtype: Link Rect: 251, 629, 258, 641 ActionDest: 101 Annotation: Subtype: Link Rect: 196, 308, 203, 320 ActionDest: 112 Annotation: Subtype: Link Rect: 200, 232, 208, 244 ActionDest: 171 Page: Sequence: 6 Page: Sequence: 7 Annotations: Annotation: Subtype: Link Rect: 401, 227, 408, 239 ActionDest: 223 Page: Sequence: 8 Annotations: Annotation: Subtype: Link Rect: 251, 247, 263, 256 ActionDest: 596 Annotation: Subtype: Link Rect: 362, 221, 374, 231 ActionDest: 596 Annotation: Subtype: Link Rect: 257, 159, 264, 168 ActionDest: 593 Annotation: Subtype: Link Rect: 264, 159, 276, 168 ActionDest: 596 Page: Sequence: 9 Annotations: Annotation: Subtype: Link Rect: 212, 455, 224, 464 ActionDest: 596 Annotation: Subtype: Link Rect: 303, 306, 310, 318 ActionDest: 259 Annotation: Subtype: Link Rect: 212, 85, 219, 97 ActionDest: 304 Annotation: Subtype: Link Rect: 527, 88, 539, 97 ActionDest: 596 Page: Sequence: 10 Annotations: Annotation: Subtype: Link Rect: 391, 350, 403, 359 ActionDest: 596 Annotation: Subtype: Link Rect: 441, 347, 448, 359 ActionDest: 304 Page: Sequence: 11 Annotations: Annotation: Subtype: Link Rect: 347, 720, 354, 732 ActionDest: 320 Annotation: Subtype: Link Rect: 473, 245, 480, 254 ActionDest: 352 Page: Sequence: 12 Annotations: Annotation: Subtype: Link Rect: 191, 438, 203, 447 ActionDest: 596 Annotation: Subtype: Link Rect: 384, 154, 396, 164 ActionDest: 596 Annotation: Subtype: Link Rect: 396, 154, 408, 164 ActionDest: 596 Page: Sequence: 13 Annotations: Annotation: Subtype: Link Rect: 244, 720, 251, 732 ActionDest: 355 Annotation: Subtype: Link Rect: 182, 710, 194, 719 ActionDest: 596 Annotation: Subtype: Link Rect: 331, 452, 338, 464 ActionDest: 355 Annotation: Subtype: Link Rect: 384, 212, 391, 224 ActionDest: 355 Page: Sequence: 14 Annotations: Annotation: Subtype: Link Rect: 282, 485, 294, 494 ActionDest: 596 Annotation: Subtype: Link Rect: 295, 485, 307, 494 ActionDest: 596 Annotation: Subtype: Link Rect: 438, 334, 450, 342 ActionDest: 599 Annotation: Subtype: Link Rect: 191, 167, 198, 179 ActionDest: 406 Annotation: Subtype: Link Rect: 182, 157, 194, 166 ActionDest: 596 Page: Sequence: 15 Annotations: Annotation: Subtype: Link Rect: 336, 457, 343, 469 ActionDest: 406 Page: Sequence: 16 Annotations: Annotation: Subtype: Link Rect: 218, 695, 231, 707 ActionDest: 443 Annotation: Subtype: Link Rect: 214, 367, 221, 379 ActionDest: 443 Annotation: Subtype: Link Rect: 538, 329, 545, 341 ActionDest: 352 Annotation: Subtype: Link Rect: 492, 82, 506, 94 ActionDest: 406 Page: Sequence: 17 Annotations: Annotation: Subtype: Link Rect: 217, 657, 229, 669 ActionDest: 465 Annotation: Subtype: Link Rect: 217, 209, 228, 221 ActionDest: 511 Annotation: Subtype: Link Rect: 217, 171, 229, 183 ActionDest: 511 Annotation: Subtype: Link Rect: 195, 159, 207, 171 ActionDest: 511 Page: Sequence: 18 Annotations: Annotation: Subtype: Link Rect: 377, 342, 389, 351 ActionDest: 599 Page: Sequence: 19 Annotations: Annotation: Subtype: Link Rect: 165, 647, 411, 658 Annotation: Subtype: Link Rect: 168, 611, 426, 622 Annotation: Subtype: Link Rect: 239, 319, 277, 330 Annotation: Subtype: Link Rect: 119, 297, 158, 307 Annotation: Subtype: Link Rect: 59, 251, 96, 261 Annotation: Subtype: Link Rect: 261, 204, 299, 215 Annotation: Subtype: Link Rect: 108, 182, 146, 192 Annotation: Subtype: Link Rect: 131, 159, 168, 169 Annotation: Subtype: Link Rect: 59, 136, 96, 146 Annotation: Subtype: Link Rect: 241, 89, 278, 100 Annotation: Subtype: Link Rect: 376, 66, 414, 77 Page: Sequence: 20 Annotations: Annotation: Subtype: Link Rect: 202, 676, 240, 687 Annotation: Subtype: Link Rect: 444, 652, 482, 664 Annotation: Subtype: Link Rect: 217, 630, 254, 640 Annotation: Subtype: Link Rect: 208, 607, 245, 617 Annotation: Subtype: Link Rect: 111, 549, 148, 560 Annotation: Subtype: Link Rect: 239, 526, 276, 537 Annotation: Subtype: Link Rect: 143, 504, 181, 514 Annotation: Subtype: Link Rect: 187, 504, 223, 514 Annotation: Subtype: Link Rect: 246, 480, 283, 491 Annotation: Subtype: Link Rect: 289, 480, 325, 491 Annotation: Subtype: Link Rect: 252, 458, 289, 468 Annotation: Subtype: Link Rect: 208, 399, 245, 410 Annotation: Subtype: Link Rect: 251, 399, 288, 410 Annotation: Subtype: Link Rect: 113, 377, 151, 387 Annotation: Subtype: Link Rect: 59, 343, 96, 353 Annotation: Subtype: Link Rect: 467, 319, 505, 330 Annotation: Subtype: Link Rect: 118, 297, 155, 307 Annotation: Subtype: Link Rect: 428, 192, 465, 203 Annotation: Subtype: Link Rect: 472, 192, 508, 203 Annotation: Subtype: Link Rect: 59, 159, 96, 169 Annotation: Subtype: Link Rect: 59, 90, 95, 100 Page: Sequence: 21 Annotations: Annotation: Subtype: Link Rect: 373, 687, 410, 698 Checksum: dc66315f Type: CRC32 Checksum: c8b19c8928f17be749658fdf1db06733 Type: MD5 Checksum: 192a012e7ba333fc18701afac59c243bf9c944c4 Type: SHA-1