Soft Sensing of LPG Processes Using Deep Learning.

This study investigates the integration of soft sensors and deep learning in the oil-refinery industry to improve monitoring efficiency and predictive accuracy in complex industrial processes, particularly de-ethanization and debutanization. Soft sensor models were developed to estimate critical variables such as the C2 and C5 contents in liquefied petroleum gas (LPG) after distillation and the energy consumption of distillation columns. The refinery's LPG purification process relies on periodic sampling and laboratory analysis to maintain product specifications. The models were tested using data from actual refinery operations, addressing challenges such as scalability and handling dirty data. Two deep learning models, an artificial neural network (ANN) soft sensor model and an ensemble random forest regressor (RFR) model, were developed. This study emphasizes model interpretability and the potential for real-time updating or online learning. The study also proposes a comprehensive, iterative solution for predicting and optimizing component concentrations within a dual-column distillation system, highlighting its high applicability and potential for replication in similar industrial scenarios.

Differences of apathy perfusion correlates between Alzheimer's disease and frontotemporal dementia. A 99mTc-HMPAO SPECT study with automated Brodmann areas analysis.

OBJECTIVES: To explore differences of apathy perfusion correlates between Alzheimer's disease (AD) and Frontotemporal dementia (FTD) using perfusion SPECT. METHODS: We studied 75 FTD and 66 AD patients. We evaluated apathy using Neuropsychiatric Inventory (NPI). We compared perfusion of BAs on left (L) and right (R) hemisphere in AD and FTD. RESULTS: Apathy in AD was significantly and negatively correlated with dorsolateral prefrontal cortex bilaterally, right anterior prefrontal cortex, inferior frontal cortex bilaterally, especially on the right, orbital part of inferior frontal gyrus bilaterally, left dorsal anterior cingulate cortex, right primary and secondary visual cortex, and with bilateral anterior and dorsolateral prefrontal cortex, inferior frontal cortex and orbital part of inferior frontal gyrus, bilaterally, bilateral anterior -ventral and dorsal- cingulate cortex, left posterior ventral cingulate cortex, right inferior, middle and anterior temporal gyri, entorhinal and parahippocampal cortex in FTD. CONCLUSIONS: Significant overlapping of apathy perfusion correlates between AD and FTD is seen in frontal areas and anterior cingulate. Right occipital cortex is also involved in AD, while right temporal cortex and left posterior cingulate are involved in FTD. Nuclear imaging could be a useful biomarker for revealing apathy underlying mechanisms, resulting in directed treatments.KEYPOINTSUnderlying neural networks and clinical manifestation of apathy may differ between AD and FTD.Apathy in AD is correlated with hypoperfusion in bilateral frontal areas, more prominent on the right, left anterior cingulate and right occipital cortex.Apathy in FTD is correlated with hypoperfusion in bilateral frontal areas, bilateral anterior cingulate, left posterior cingulate and right temporal cortex.Brain perfusion SPECT with automated BAs analysis and comparison with normal healthy subjects may provide significant information for apathy mechanisms in neurodegenerative disorders, affecting patients' treatment.

Anosognosia in Dementia: Evaluation of Perfusion Correlates Using 99mTc-HMPAO SPECT and Automated Brodmann Areas Analysis.

(1) Background: Considerable inconsistency exists regarding the neural substrates of anosognosia in dementia in previous neuroimaging studies. The purpose of this study was the evaluation of anosognosia perfusion correlates across various types of dementia using automated Brodmann areas (BAs) analysis and comparison with a database of normal subjects. (2) Methods: We studied 72 patients: 32 with Alzheimer's disease, 26 with frontotemporal dementia-FTD (12 behavioral FTD, 9 semantic FTD, 5 Progressive Non-Fluent Aphasia), 11 with corticobasal syndrome, and 3 with progressive supranuclear palsy. Addenbrook's Cognitive Examination-Revised (ACE-R) mean(±SD) was 55.6(±22.8). For anosognosia measurement, the Anosognosia Questionnaire-Dementia was used. Total anosognosia score mean(±SD) was 22.1(±17.9), cognitive anosognosia score mean(±SD) was 18.1(±15.1) and behavioral-mood anosognosia score mean(±SD) was 3.3(±4.7). (3) Results: Higher anosognosia total score was associated with hypoperfusion in the inferior temporal, anterior cingulate, and inferior frontal cortices of the right hemisphere (BAs 20R, 24R, 32R, 45R). Higher anosognosia cognitive score was correlated with hypoperfusion in the left middle and anterior temporal cortices, and right dorsal anterior cingulate cortex (BAs 21L, 22L, 32R). No association was found with behavioral-mood anosognosia. (4) Conclusions: Automated analysis of brain perfusion Single Photon Emission Computed Tomography could be useful for the investigation of anosognosia neural correlates in dementia.

Correlation of Neuropsychiatric Symptoms in Dementia with Brain Perfusion: A 99mTc-SPECT-HMPAO Study with Brodmann Areas Analysis.

BACKGROUND: Neuropsychiatric symptoms (NPSs) are common in dementia. Their evaluation is based on Neuropsychiatric Inventory (NPI). Neuroimaging studies have tried to elucidate the underlying neural circuits either in isolated NPSs or in specific forms of dementia. OBJECTIVE: The objective of this study is to evaluate the correlation of NPS in the NPI with Brodmann areas (BAs) perfusion, for revealing BAs involved in the pathogenesis of NPSs in dementia of various etiologies. METHODS: We studied 201 patients (82 with Alzheimer's disease, 75 with Frontotemporal dementia, 27 with Corticobasal Syndrome, 17 with Parkinson Disease/Lewy Body Dementia). Exploratory factor analysis was carried out to evaluate underlying groups of BAs, and Principal Component analysis was chosen as extraction method using Varimax rotation. Partial correlation coefficients were computed to explore the association of factors obtained from analysis and NPI items controlling for age, educational yeas, and ACE-R. RESULTS: We found 6 BAs Factors(F); F1 (BAs 8,9,10,11,24,32,44,45,46,47, bilaterally), F2 (BAs 4,5,6,7,23,31, bilaterally), F3 (BAs 19,21,22,37,39,40, bilaterally), F4 (BAs 20,28,36,38, bilaterally), F5 (BAs 25, bilaterally) and F6 (BAs 17,18, bilaterally). Significant and negative correlation was found between NPI1 (delusions) and F3,F6, NPI2 (hallucinations) and F6, NPI7 (apathy) and F1,F4,F5, NPI3 (agitation) - NPI10 (aberrant motor behavior) - NPI12 (eating disorders) and F1. We did not find any significant correlation for NPI4,5,6,8,9,11 (depression, anxiety, euphoria, disinhibition, irritability, sleep disorders, respectively). CONCLUSION: Several NPSs share the same BAs among different types of dementia, while the manifestation of the rest may be attributed to different neural networks. These findings may have an impact on patients' treatment.

Eating Disorders in Frontotemporal Dementia and Alzheimer's Disease: Evaluation of Brain Perfusion Correlates Using 99mTc-HMPAO SPECT with Brodmann Areas Analysis.

BACKGROUND: Eating disorders (ED) in dementia represent a significant impairment affecting patients' and caregivers' lives. In frontotemporal dementia (FTD), ED include overeating, sweet food preference, stereotypical eating, and hyperorality, while in Alzheimer's disease (AD), anorexia and appetite loss are the most common ED. OBJECTIVE: The aim of our study was to highlight Brodmann areas (BAs) implicated specifically in the appearance of ED in FTD and AD. METHODS: We studied 141 patients, 75 with FTD and 66 with AD. We used the NeuroGamTM software on the reconstructed single photon emission computed tomography-SPECT data for the automated comparison of BAs perfusion on the left (L) and right (R) hemisphere with perfusion in corresponding BAs of a normal database. RESULTS: The FTD group included 27 men and 48 women, age (mean±SD) 65.8±8.5 years, duration of disease 3.4±3.3 years, Mini-Mental State Examination (MMSE) 17.9±8.6, ED score on Neuropsychiatric Inventory (NPI) 4.7±8.5. ED in FTD were correlated with hypoperfusion in right anterior and dorsolateral prefrontal cortices (BAs 10R, 46R), left orbitofrontal cortex (BA 12L), orbital part of the right inferior frontal gyrus (BA 47R), and left parahippocampal gyrus (BA 36L). The AD group included 21 men and 45 women, age (mean±SD) 70.2±8.0 years, duration of disease 3.3±2.4 years, MMSE 20.2±6, ED-NPI score 2.7±3.9. ED in AD were correlated with hypoperfusion in left inferior temporal cortex (BA 20L). CONCLUSION: SPECT imaging with automated mapping of brain cortex could contribute to the understanding of the neural networks involved in the manifestation of ED in dementia.

Clinical Evaluation of Brain Perfusion SPECT with Brodmann Areas Mapping in Early Diagnosis of Alzheimer's Disease.

Early diagnosis of Alzheimer's disease (AD) based on clinical criteria alone may be problematic, while current and future treatments should be administered earlier in order to be more effective. Thus, various disease biomarkers could be used for early detection of AD. We evaluated brain perfusion with 99mTc-HMPAO single photon emission computed tomography (SPECT) and Brodmann areas (BAs) mapping in mild AD using an automated software (NeuroGam) for the semi-quantitative evaluation of perfusion in BAs and the comparison with the software's normal database. We studied 34 consecutive patients with mild AD: 9 men, 25 women, mean age 70.9 ± 8.1 years, mean Mini-Mental State Examination 22.6 ± 2.5. BAs 25L, 25R, 38L, 38R, 28L, 28R, 36L, and 36R had the lower mean perfusion values, while BAs 31L, 31R, 19R, 18L, 18R, 17L, and 17R had the higher mean values. Compared with healthy subjects of the same age, perfusion values in BAs 25L, 25R, 28R, 28L, 36L, and 36R had the greatest deviations from the healthy sample, while the lowest deviations were found in BAs 32L, 32R, 19R, 24L, 17L, 17R, 18L, and 18R. A percentage of ≥94% of patients had perfusion values more than -2SDs below the mean of healthy subjects in BAs 38R, 38L, 36L, 36R, 23L, 23R, 22L, 44L, 28L, 28R, 25L, and 25R. The corresponding proportion was less than 38% for BAs 11L, 19R, 32L, 32R, 18L, 18R, 24L, and 17R. In conclusion, brain SPECT studies with automated perfusion mapping could be useful as an ancillary tool in daily practice, revealing perfusion impairments in early AD.

Brain perfusion SPECT with Brodmann areas analysis in differentiating frontotemporal dementia subtypes.

Despite the known validity of clinical diagnostic criteria, significant overlap of clinical symptoms between Frontotemporal dementia (FTD) subtypes exists in several cases, resulting in great uncertainty of the diagnostic boundaries. We evaluated the perfusion between FTD subtypes using brain perfusion (99m)Tc-HMPAO SPECT with Brodmann areas (BA) mapping. NeuroGam software was applied on single photon emission computed tomographic (SPECT) studies for the semi-quantitative evaluation of perfusion in BA and the comparison with the software's normal database. We studied 91 consecutive FTD patients: 21 with behavioural variants (bvFTD), 39 with language variants (lvFTD) [12 with progressive non-fluent aphasia (PNFA), 27 with semantic dementia (SD)], and 31 patients with progressive supranuclear palsy (PSP)/corticobasal degeneration (CBD). Stepwise logistic regression analyses showed that the BA 28L and 32R could independently differentiate bvFTD from lvFTD, while the BA 8R and 25R could discriminate bvFTD from SD and PNFA, respectively. Additionally, BA 7R and 32R were found to discriminate bvFTD from CBD/PSP. The only BA that could differentiate SD from PNFA was 6L. BA 6R and 20L were found to independently differentiate CBD/PSP from lvFTD. Moreover, BA 20L and 22R could discriminate CBD/PSP from PNFA, while BA 6R, 20L and 45R were found to independently discriminate CBD/PSP from SD. Brain perfusion SPECT with BA mapping can be a useful additional tool in differentiating FTD variants by improving the definition of brain areas that are specifically implicated, resulting in a more accurate differential diagnosis in atypical or uncertain forms of FTD.

Perfusion SPECT studies with mapping of Brodmann areas in differentiating Alzheimer's disease from frontotemporal degeneration syndromes.

OBJECTIVES: The aim of this study was to evaluate the contribution of brain perfusion single-photon emission computed tomography (SPECT) studies with mapping of Brodmann areas (BAs) in the differential diagnosis between Alzheimer's disease (AD) and frontotemporal degeneration (FTLD) syndromes. METHODS: Thirty-nine patients with AD and 73 patients with FTLD syndromes [behavioural variant FTLD (bvFTLD); language variant FTLD (lvFTLD), including semantic dementia (SD) and progressive nonfluent aphasia (PNFA); and corticobasal degeneration (CBD)/progressive supranuclear palsy (PSP) syndromes] underwent brain perfusion SPECT. The NeuroGam software was used for the semiquantitative evaluation of perfusion in BAs of the left (L) and right (R) hemispheres. RESULTS: Compared with those in AD patients, BAs with statistically significant hypoperfusion were found in the prefrontal, orbitofrontal and cingulated cortices and Broca's areas of FTLD and bvFTLD patients; in the temporal and prefrontal cortices and Broca's areas of lvFTLD patients; in the left temporal gyrus of SD patients; in premotor and supplementary motor, prefrontal, orbitofrontal, temporal and anterior cingulated cortices and Broca's areas of PNFA patients; and in the prefrontal, temporal, posterior cingulated and primary and secondary visual cortices of CBD/PSP patients. BA 46R could differentiate AD patients from FTLD and bvFTLD patients; 21L and 25L were found to be independent predictors for lvFTLD in comparison with AD, and 25R, 21L and 23R could differentiate AD patients from PNFA, SD and CBD/PSP patients, respectively. CONCLUSION: Brain perfusion SPECT with BA mapping in AD and FTLD patients could improve the definition of brain areas that are specifically implicated in these disorders, resulting in a more accurate differential diagnosis.

Radiopharmaceuticals in neurological and psychiatric disorders.

The development of functional brain nuclear medicine techniques and their application in the investigation of neuropsychiatric disorders, have contributed significantly in the illumination of the underlying pathophysiological processes of these disorders. Furthermore, Single Photon Emission Computed Tomography (SPECT) and Positron Emission Tomography (PET) brain studies provide information in early diagnosis, differential diagnosis, development of new drugs, and monitoring the response to therapeutic management. SPECT and PET brain imaging require the use of radiopharmaceuticals that cross the intact Blood Brain Barrier (BBB). Such radiotracers have been used in regional Cerebral Blood Flow (rCBF) SPECT and PET imaging and brain metabolism imaging with PET; these are well established methods in the diagnosis and management of various cerebral vascular diseases (e.g. stroke, dementia, epilepsy). Advances in radiotracer chemistry have resulted in the development of molecular imaging which represents the molecular and cellular processes of neuropsychiatric diseases. SPECT and PET molecular imaging has become available for the study of acetylcholinergic, dopaminergic and serotonergic systems, as well as for benzodiazepine and opioid receptors, with promising results. More studies are needed to validate the role of molecular imaging in the clinical practice of neuropsychiatric disorders.