Early detection of neurodegenerative diseases, such as Alzheimers and Parkinsons, is essential for timely intervention, which can improve patients' quality of life and slow down disease progression. Traditional diagnostic methods rely heavily on clinical tests, which can be infrequent and may not capture slight behavioral changes that indicate early cognitive or motor decline. This work presents a novel approach using smart home data to detect early signs of neurodegeneration through continuous monitoring of sleep patterns and daily activity routines. In a smart home environment, sensors passively monitor daily routines, sleep quality, and mobility patterns of the elderly persons. This paper introduces a novel framework combining the Sleep Deviation Patterns (SDP) and the Weighted Activity Deviation Index (WADI) to comprehensively analyze deviations in sleep and daily routines. The SDP framework captures deviations in sleep onset, duration, interruptions, and consistency using metrics such as Sleep Onset Deviation, Sleep Duration Deviation, Sleep Interruption Index, and Sleep Consistency Index-aggregated into a weighted Sleep Deviation Score. WADI quantifies deviations in daily activities by computing weighted absolute deviations of activity proportions relative to a reference routine. Thus, we applied our framework to real-world smart home datasets (TM001-TM004) from the CASAS project, which include labeled activity data from both single- and multi-resident households. Experimental findings reveal a distinct stratification: TM001 and TM002 exhibited Low WADI and Low SDP with an average of >0.015 - >0.2 scores, suggesting stable routines, whereas TM003 and TM004 demonstrated elevated with an average of >0.03 - >0.4 values, indicating disrupted behaviors. In TM004, up to 28% of days were flagged as anomalous, correlating with patterns consistent with early neurodegeneration such as fragmented sleep and disorganized activity routines. Finally, experimental results demonstrate that the combined SDP and WADI frameworks effectively identify irregularities in sleep and activity patterns on real-world datasets. The proposed approach offers a robust and scalable solution for health monitoring, with potential applications in neurodegenerative disease detection, personalized healthcare, and smart home systems.