Can someone please explain how the data integrity in the Merkle Patricia Trie is verified?

If you're a light client (SPV), I understand how in the regular Merkle Tree scenario in bitcoin you simply concatenate the hashes and rehashing them upward in the path toward the root USING with the Merkle Proof provided by the full node. But how is this done in Ethereum? Cause from what I understand about the Merkle Patricie Trie, it's different from the Merkle Tree in that you can't verify construct/calculate it by concatenating hashes and rehashing them. What exactly is the process used here to calculate the hashes upward eventually leading to the calculation of the root hash?